TL; DR for ISYE 6501: Interesting lecture material + homework (15%) and project (8%) with work potentially sabotaged by peer grading. Horrible exams (75%). Some of the exam questions were so oddly worded that if English was not your native language you were at a distinct disadvantage. Enrollment: 1300+ students! Class subject matter+ HW assignments + project: 5/5 Peer grading and exams: 1/5 Overall course rating: 3/5. Exams and peer grading are red flags.

Grades: HW + Project 100, Exams (average) 87, Final grade A (90%)

Pros (Top 3):

1. Excellent choice of topics and covered in just enough detail by Dr. Sokol in well-organized lectures.
2. Interesting homework assignments, designed with the lectures in mind. HW consists of 80% of your time in this class and where you will put work in. This course needs to use the auto-grader for the coding.
3. The project assignment allows a deep dive into a particular problem and requires you use at least three of the methods acquired in the first 10 weeks. As with the homework you get out of the project what you put in.

Cons (Bottom 3):

1. Exams: 75% of Grade. It’s a shame that the often ambiguous and poorly worded exams account for 75% of the grade in this course. Many exam questions introduce an unnecessary level of ambiguity that is unsettling.
2. HW: Homework and Project (23% of Grade) are all peer graded, and account for almost a quarter of the grade. Some of the peer reviews I received were performed with only a modicum of effort if any. Given the quality of gatech.edu as an engineering school, the lack of an auto-grader for coding exercises is unforgivable.
3. Office Hours (OH): Due to work pressures I had to finish the HW well before OH on Monday. So, OH were not very useful. As reported in OMS reviews, some students waited for the office hours the Monday before the HW was due and then copied the work done there. Some of the HW rubrics were weak and nearly incomplete.

Six course design fixes that would allow ISYE 6501 to achieve a 5/5 rating:

1. Exams: Make exams better reflect the material not the ability to dissect triple negatives and split infinitives.
2. Grading: Please reward work on homework writeups. At minimum a 50%-50% split between HW/Project and Exams.
3. Coding: Include a 5-10-minute TA-led segment demonstrating R or python code relevant to HW as part of the lecture.
4. HW: Implementation should require the student to deep dive the details. Provide more test problems.
5. Require 2+ of the HWs (chosen randomly) to have one TA grade to ensure quality grading & identify poor peer graders.
6. Get rid of Piazza, use Ed.

I wanted to share some honest feedback after taking two midterms of ISYE 6501. I’ve really enjoyed the lectures and learned a lot from the assignments — even though they’re quite time-consuming, they’ve been deeply educational and practical. But I feel that the exam setup and grading structure don’t really align with the spirit or learning goals of the course.

To start with, the exam questions feel misleading and unevenly weighted. Some questions are worth 5–6 points while others are only 1 point, and it doesn’t seem that this distribution was carefully reviewed. In many cases, it felt like the challenge was not about applying what we learned, but about trying to interpret what the question writer was actually asking. That creates unnecessary stress and confusion, unrelated to the actual course material.

I can’t help but wonder — are the exams thoroughly reviewed by Prof before release? Because the quality of the lectures is excellent, but the exam feels disconnected from that level of care.

Also, I think the grading structure overall needs reconsideration. Homework only counts for 15% of the grade, yet it easily represents 60–70% of the actual learning in this course. I understand it might be weighted lower because it’s peer-graded and perhaps not taken as seriously in scoring, but that’s exactly the problem — it’s where most of the effort and meaningful understanding happen. If that’s the case, then the peer-grading system itself should be reevaluated, not just for the sake of TA workload, but to ensure that students are fairly rewarded for the depth of learning these assignments provide.

Meanwhile, exams make up 75% of the grade, even though they often don’t reflect how well someone understands the material — just how well they can navigate ambiguous or oddly worded questions.

I truly think ISYE 6501 is a great course with strong fundamentals, but it could be so much better if the exams and grade weighting were revisited to better reflect the effort and real learning happening throughout the semester.

Fantastic introduction to the program that provides a broad survey of analytical methods. The difficulty will be mostly dependent on your background with coding, R programming, and data analytics in general. Dr. Sokol's videos are fun and easy to follow. The office hours go over homework problems. It was still a great refresher course even with years of coding and data experience. I would highly recommend taking this as your first course!

This class is a good introduction to analytic models. The homework can be tedious if you are unfamiliar with R. Most of my time for this class was spent on the homework. The lectures are well made and easy to follow. The exams, while much harder than the practice quiz, are fair. They tend to test you on your understanding of the concepts, thus no memorization is required. You are also allowed cheat sheets for the exams, which helps solidify your understanding.

I had mixed feelings about this course.

On the positive side, I found the course content and rigour to be a solid way to kickstart my OMSA journey. The professor is super engaging and enthusiastic, and his lectures explain both the theory and examples very well. That said, the homework is largely self-directed - you’re expected to code independently (+ learn to code for those with no R background). The first homework was by far the most challenging for me (and possibly for other students too), which caught me off guard. After that, the remaining assignments were more manageable. Just note that there are one or two of the later weeks that need Python, which can be tricky if you don’t have prior experience.

Peer grading was frustrating at times since it’s inherently subjective - often the difference between a 90% and 100% depends on the grader, and many seemed to approach it as a chore rather than a peer learning opportunity. Personally, I would have preferred more (or any) TA grading, though given that homework makes up only a small portion of the final grade, it’s hard to complain much. There’s also a mini project, but it’s fairly light - no coding required, just a short report of your problem/use cases and proposed analytics.

The quizzes are what really determine your final grade. As long as you keep up with the lectures, you should be fine, though some of the questions are worded in a tricky way. Unfortunately, the quizzes don’t tie in with the homework at all, which further adds to my impression that the course workload feels disjointed.

TLDR: Great professor and lectures, but homework doesn't really align with the effort required (so don’t stress over it too much).

As an OMSCS student this was a great course to take and is basically a statistical learning / ML 101 course. It doesn't dive too deep into the math behind any models but will build a good foundation of how to approach modeling and what models are appropriate for what scenarios.

The lectures were a 10/10 and extremely informative, the homework could take a while but were really good for solidifying information. Although homework was only 15% of the grade and is peer reviewed (a 90 is for a complete homework, 100 is only if you go above and beyond). I found that even if I went above and beyond sometimes I'd get graded a 90 and other times a 100 so it's a toss up.

The course is exam heavy (75% of your grade across 3 exams) and 95% of the information is from the lectures. The median grades were low to mid 80s and I found t exams to be fairly challenging. Some of the wording for the questions threw me off but you get used to it after the first exam.

My advice would be to just do the homework to completion and put the rest of your effort into learning the material for the exams. There also seems to be a small unofficial curve (I got an A with an 89% actual grade). This course is an easy B if you do the work and have decent exam grades. To get an A you really need to crush the exams or get an 87+ and hope for a curve.

Overall it was a great course to take, just be prepared to get a B if you struggle on the exams.

I liked the course but it's not as easy as some people make it out to be. Professor Sokol is very knowledgeable and following the lecture content is rarely dry. The videos are well-made and the material is important for your analytics career as you cover important topics like regression, principal component analysis, and optimization to name a few. This is an introductory course so don't expect a full in-depth review. It's just enough to get your feet wet.

Homeworks (15%), 1 project (8%), and intro survey (2%) take 25% of grade while the remaining 3 exams take 75%. You will need to learn R for the homework's which can be time consuming in the beginning. 2 homework's are dropped so do well on the earlier assignments. They are graded by your peers so it's pretty difficult to get 100's unless you go above and beyond which for 15% doesn't really seem worth it. I ended up doing well (mostly 90s on the beginning homework's) and just didn't do the last homework's regarding optimization which was the last topic of the course. The project is just a super simple report on an analytics case study you choose (no coding required).

Now we get to the exams. I'm generally a decent test taker but these exams are worded in such an unusual way that it almost throws you off guard. Not even the content of the questions but the phrasing. Multiple answers may fit but the best answer is needed which can be confusing well. I think the mean for all the exams was high 70s, low 80's. The exams are my only gripe with the course, I wish they were a little more traditional. The homework's are not needed for the exams so in theory you could ignore every single homework and still pass if you do well on the exams (but don't do this lol).

Disclaimer: the class also uses Piazza (I think all the OMSA use it but I could be mistaken) and I rarely went on it.

I'm in the CS track and I'm glad i took this course. One of the better courses to take given the importance of its content and how it was presented just don't expect an easy A. An easy C, a B with decent effort, and an A if you do really well on the exams.

The Good: The material was interesting and the lectures were pretty good. With a background in analytics, the material was not extremely difficult but still provided some challenge at points. There are plenty of TAs, most of whom were extremely helpful. Most Piazza posts were answered promptly and well. The office hours were extremely helpful in completing the homework assignments.

The Bad: The structure of the course is a little weird at points, but my major gripe with this class was the exams. The homeworks do absolutely nothing to prepare you for the exams, so don't bother with reviewing those to study. They are also worth so little of your grade compared to the amount of time you spend on them that it almost feels like a waste of time. The exams make up 75% of your grade, which wouldn't be as big of a deal if the exams were written better. Some of the question wording in the exams is intentionally confusing, trying to force you to make a mistake. I never understood why exams would contain "Gotcha" questions that don't actually test your understanding of the material. The second exam was the most clearly written, but midterm 1 and the final were more confusing. The first two exams contain essentially no questions regarding code, but the final randomly has questions about R functions. They also like to have true/false questions that aren't formatted as true/false (you are given a statement with a missing word and have to choose from 2 options like COULD/COULD NOT). Additionally, the question count on the exam is a little misleading as they love to have several questions with 6+ parts to them, especially in the final. I also felt that Dr. Sokol was relatively absent from the courses as it's clear the lectures were recorded years ago and he does not attend most if any of the office hours. Another major gripe is having all assignments and the project be peer graded. This leads to a lot of inconsistency in the grades for these. The grading scale is a bit annoying (getting everything correct only gets you a 90%, to get a 100% you are supposed to provide a "Deeper solution than expected". Some weeks I put in a ton of effort and got a 90%, other weeks I put in very little effort and still got a 90% or one time an 100%. Some students oddly get a bit of a power trip on the peer grading and nitpick the tiniest things to take points away (which is honestly a little sad). One student gave me a 75% on an assignment saying I got everything right but they didn't like the way I formatted my assignment.

Overall, it's a fine class, but I wish it was better especially given it is a core class. I felt like most of the reviews I read pretty well summed up the course. I think it could benefit from having a bit more project based work, with something similar to CSE 6040 having autograded coding assignments and exams.

I came into this course with very little R experience or knowledge, so the ramp in the beginning was pretty steep. While the initial shock softens after 2-3 weeks, the class is by no means easy, but the effort and time are worth the volume of useful information you learn. If you want to make the most of the course, use the ISLR as a supplement and make the most of the homework assignments - experiment & go beyond the assignment. It's worth it.

I've read some reviews complaining about the wording on test questions. I think they're mostly correct: some of the questions do seem to be worded/structured to create confusion. Still - it's Georgia Tech. Every class you take is going to try to make sure you're thinking, learning and applying the info, not just regurgitating facts. They don't just give away Master's degrees.

Really good class, low workload if you're already proficient at R. The lectures are fantastic and the homeworks are pretty thought-provoking. I spent like an hour on lectures+notes and random Googling each week, maybe 2 hours on homework, and then studied more in the weeks leading up to exams.

I wish the grade weighting was done a little differently. 75% comes from some pretty challenging exams with questionable wording, whereas homework is just 15% and is very time-consuming.

The course project also left a lot to be desired. I thought we were going to get to do a real analytics project, but instead it's basically just a write-up.

Advice for future students: make sure you know your stats/math. If you don't know R, learn it before the course - this will save you a ton of time during the semester. Take good notes for your cheat sheet as the weeks go by so you're not scrambling to make a cheat sheet last minute. Take your time on exams and read every question carefully. Study hard for the exams since they are most of your grade.

Taking ISYE 6501 as my first course in Georgia Tech’s OMSA program was an excellent decision. It provided a broad yet thorough introduction to analytics modeling, balancing theoretical knowledge with practical application. Dr. Sokol’s engaging teaching style, the challenging but rewarding homework, and the valuable support resources like TA office hours and the Stats Bootcamp made the learning experience both enjoyable and effective. If you’re considering the OMSA program or simply want to strengthen your analytics foundation, I highly recommend ISYE 6501. It’s a course that delivers real value and sets the stage for success in the field of analytics.

I've written a more detailed review and course summary here: https://blog.marketingdatascience.ai/introduction-to-analytics-modeling-my-review-of-georgia-tech-isye-6501-4d35f5b0482b

This is exactly what an intro survey course should be.

Workload will vary a lot by person. I took a long time on assignments to try to get 100s, which really enriched my understanding of the material and required me to explain each model thoroughly. The first test had some tricky questions, but the two others were somewhat straightforward IF you thoroughly understand the lecture material.

I really liked this class and appreciated Dr. Sokol's lectures. He does a great job explaining enough so that you understand the underlying reasons why a model does what it does without getting so in the weeds that it becomes an LA or Calc class. A large group of very helpful TAs also made this a great class.

I took a short R programming class on Udemy before taking this class. I have almost no prior programming experience, and learning beforehand really helped. I got a 100 on all assignments and mid 90s on all exams.

Overall, I thought this was a great course with a great instructor and group of TA's, but I wouldn't call it an easy A by any means.

The lecture videos are some of the best, if not the best, in the program. Joel Sokol is a natural at explaining concepts in his videos in a way that simplifies what may otherwise be complex topics in a way that makes them understandable, which is perfect for an intro/survey course. The TA's in piazza are also incredibly helpful and responsive when it comes to answering questions you throw their way. These were definitely highlights of the course.

The homework difficulty and time commitment is dependent on your experience in R. If you're new to R, you may struggle initially, but it gets better as you get used to it. Worst case scenario, the TA's provide starter code in office hours shortly before the assignments are due which should help you figure out what needs to be done on the assignment to get either a 90 or 100 on them.

What I found to be difficult in this course was the pacing of it. I don't recommend taking it in the summer as the class felt pretty non stop even when I took it in the fall. Every week there's new material being taught, with a homework due, and then you have to complete three peer reviews. This is the same case for exam weeks: new material and homework is released even as you're trying to prepare for the exam which I found kind of rough. As mentioned before, you'll learn a lot and it's well taught, so this is just something to be aware of.

Additionally, the tests themselves aren't the easiest and they make up a large portion of your grade. The exams are purely conceptual and come straight from the lectures, so don't refer to homework to prepare for these. The way some of the questions are worded can get tricky and may require you to think pretty heavily, so my biggest recommendation is to really listen to the videos closely and make sure you really understand everything fully. I was able to do well through this approach but it definitely was time consuming, and I understand why people complain about the exams.

Again it's a great course and you'll learn a ton, but the pacing is fast and the exams require a very strong understanding of the material and having to think through some tricky questions

This course was challenging for me because I felt that I did not have a strong math background. I did not take linear algebra or discrete math before this class, and felt a bit lost on the formulas and whatnot. When the lecture videos touched on them, my eyes sorta just glazed over them. But they aren't necessarily heavily used in the assignments.

The coursework provided me a good overview of analytical models, what they are, when to use them, etc. The homework assignments were still doable because the TAs did a great job with the office hours in providing guidance. The assignments are peer reviewed. So I would say, if you're struggling, just do your best, attend the office hours, and submit what you can. You can still get a passable grade on the assignments I think.

There were 2 midterms and 1 final which made up a large chunk of your grade. I got mid-70s on the midterms, and mid-80s on the final. Overall, I was able to get a B in this class.

Lastly, I have to say, this course has a very supportive TA team. I've signed up for classes where the TAs can be very condescending and all, and it kills the enjoyment of the course. But this TA crew is A+ and cares about the students and how they progress. Thanks!

A great overview of analytic! Dr. Sokol's focus on learning, understanding, and real-life application of knowledge is highly appreciated. The case studies are particularly valuable for understanding real-world applications and challenges in data analytics.

Great intro course! I would recommend taking this toward the start of your degree, since a lot of the other courses dig deeper into topics introduced here. It is very much an overview of various models. You can tell Prof Sokol has a lot of successful real world experience, and a lot of his guidance is directly practical, which I really appreciated. Some of the topics lacked depth, but that sort of comes with the territory for this sort of overview course. However I strongly disliked that the homeworks and final project are peer reviewed - it's hard to get solid feedback from people who are just as inexperienced as youl

This course does not deserve its high rating. None of the lectures gave me the feeling I was actually learning anything, they were more of him throwing random examples at never explaining anything in a way that would make it intuitive. It was standard for me to have a notebook open, watch 5 lectures, and have written down nothing other then the names of the lectures.

The assignments were tedious and took upwards of 4 hours a week for a total of 15% of the grade. Keep in mind, I got 100 on every assignment, I just did them with disdain knowing that it had little bearing on my final grade. Further, I found that many of the assignments I was peer reviewing were really bad. For example, in the most recent week I had one person who clearly copied full outputs from chatGPT that barely answered the questions and another who included nothing but copy and pasted code and there outputs (no graphs, analysis, explanations, nothing).

The exams were fine but I did find some of the questions hard to understanding and misleading which was frusturating.

Taken Spring 2024: I have no idea why this class appears to be as highly regarded as it is. For the record, I got an A, and this is a terrible class.

The class is taught entirely by TAs and the professor, the director of the OMSA program, can’t be bothered to even show up for one class session. Most of the TAs provide good instruction but it’s hit or miss. Some are much better instructors than others. It’s tough luck if you get a TA with poor instructional skill because questions asked during the office hours are frequently met with “you can rewatch the video” or “you can ask that on Piazza”. Piazza is a whole other problem I’ll discuss later.

The video lectures are the primary way information is delivered and what will form the basis of the tests. Homework has nothing to do with the tests. The lectures are short and high level while the tests are detail oriented and confusingly worded. I feel sorry for anyone that doesn’t have a solid mastery of English. The class is also bizarrely organized. It starts with SVM then jumps around between supervised and unsupervised learning models. A cursory review of any intro textbook (and I have read several) will provide a logical flow of information. For example, SVM is always covered late in an intro book. This class should just use ISLR as a textbook, it’s right there.

The homework is time consuming and worth minimal to your grade. Most people seemed to spend their time learning how to code it than learning the underlying analytics techniques. Yes, a student is supposed to have “coding proficiency” for the program, but when OMSA routinely accepts people without such proficiency and tells them they can pick it up on the fly, they have implicitly altered the requirements of the program. This affects the peer review grading, which is a terrible and idiotic method of grading in grad school. For example, I had issues with RStudio crashing and was unable to complete one assignment. I went ahead and turned in what I had. I got 100 on it because two peers graded it 100 without any commentary. Another assignment received 90 as the final grade, but one peer graded it 75 because “I was interested in more explanation about the data” (which was the entirety of the comment). The data in question was “from your personal or work experience, what kinds of data might be used on a classification model?”, the HW had three more parts and no “explanation” of the data was asked for in the HW.

The point of all that, is that your HW grade will be a complete crap shoot with minimal to no reason as to why. You certainly will not receive any response that will be educational. Given that HW is also worth so low in the final grade, it’s best just to learn how to make the report look pretty and turn in the bare minimum. That will get you 90 to 100 almost every time even if you’re completely wrong. It’s much more worthwhile to spend your time understanding the lecture material in this class than doing HW.

Piazza is a terrible forum system and I have no idea why it’s being used by GT at all. There’s so many people in the class that it’s just flooded with thousands of posts. Everything from relevant questions to people that didn’t read the syllabus to questions out of left field. It was just too much to wade through.

The class is a fairly easy A, just focus on the lectures, like really understanding them, and make your HW look prettty. Then you’re all good.

How to succeed: Listen and UNDERSTAND all the lesson materials. Try your best on the HW and review the answers later. This class is not a hard class, but as the name suggests, just an introduction.

Overall, I do not feel like I wasted time, but I do not feel like I gained a lot. But I am still giving it 5 points because I got things out of the course that is exactly aligned with the name. introduction to modelling.

I wish this class was an elective for the A track guys. As I am C track, I found this course not so relevant for my future, but who knows.

Pros: subject matter was interesting, lectures were clear, TAs

The topics covered were the best part of the course. The material is interesting and highly applicable. Dr. Sokol states at the beginning that he is essentially toeing the line of giving enough information about a lot of topics without going into a complexity that goes beyond an intro course. I believe he accomplished this goal well in his lectures, as they were well explained but not too complicated. The TAs were helpful, highly responsive, and receptive to dumb or rude questions/comments.

Cons: not many opportunities for application, tricky assessments, course structure

I most of all wish there were more example (independent work) problems in the course for students to practice. There are check-in questions at the end of each module and a coding homework assignment each week. While the homeworks were helpful in helping me develop understanding of the concepts and the basics of how to implement them in code, they just weren't enough for me personally for the assessments. The course is structured moreso as a notetaking course in which you take the knowledge from the lectures and apply them straight to the test with little practice beforehand. For instance, many of the questions on the assessments ask you to apply topics in the lectures in ways that I had never practiced independently before (e.g. the modeling decisions in midterm 1 and the simulation questions on midterm 2 stand out to me). I don't see why there aren't at least some supplemental resources to help students practice more before the assessments. I often felt as if I was being assessed on several topics for the first time on the assessments, which contrasts heavily with 6040, in which there was a multitude of practice options.

The assessments have some tricky wording, but they were otherwise pretty fair. Outside of exam week, the work can essentially be done quickly - watching the lectures and then a few hours on homework (more in the beginning if you don't know R). I studied for about 5-10 hours on each assessment and averaged ~89 on all three.

Overall, though, the course isn't too difficult, so I ended with a ~90.9, which was considered an A. In some ways, though, I don't feel as if I actually deserved the A because I don't feel overall confident in my ability to apply the topics beyond a cursory level. I think the course can be improved just by adding in some more opportunities for students to practice applying their knowledge.

Pros: The course is a good one to introduce you to the analytics, it covers a lot of aspect of the analytics. The homework is not too hard and the schedule at the end of the semester is better than previous.

Cons: Need to do homework almost every week, plus the peer review, I feel that I have no time to take a breath. The percentage of the 3 exams weight too much higher (75%), you will easily drop from A to B if you did bad in one of them. I did bad in the second midterm exam but luckily I got 87.75% overall and still consider A. Overall it's a good course without too deep dive in one of the area but covers broad area.

Great introductory course to OMSA with exposure to the various part of analytics (from modelling to data cleaning / preparation). This course is focused on breadth rather than depth which is expected of an introductory course.

The lectures are there to teach the basic concepts but are no where near enough for the weekly homework assignments. The homework is the main part of the learning and is the one that takes up the most time. My only gripe is the effort-to-reward ratio for homework is low, making up only 15% of the grades while 5+ hours is spent coding + writing a report.

The 2 Midterms and Finals are weighted heavily (25% each, 75% total) where 1 slip from any of them reduce the chance of an A. I would prefer to have the weight for midterms and finals lowered and the homework to be brought up.

The TAs of this course are great with fast and immediate responses. Definitely can see that they are invested into the delivering the best experience for students.

Overall, good course for new students entering OMSA but don't expect too much depth and do factor in weekly time to do the homework assignments

The course is relatively easy, but I found the grading scheme and busy-work (homework every single week) to be archaic and not conducive to learning. I think the course is fine if it is your first/only intro to data analysis using machine learning tools in R, but if you have one previous semester of ML experience or some background in R/data analysis, it would be probably be a better use of your time to take a course that goes in more depth.

The video lectures were well taught. TA's were extremely responsive on Piazza. Exams were pretty reasonable, and I felt like the curve was decently generous.

I only wished that some of the topics had gone more in depth. But I was coming in with some modeling background, and I suppose that greater depth is what electives are for!

Overall good course. My only issue is that its ridiculous to assign 75% of the grade to 3 exams and just 15% to the homeworks, of which there are 14 and they require a long time to do. You could decide not to do any homework and still get a solid B in the class. The exams at best should count for 55 to 60% of the grade.

A very nice introduction course for new student. I took 6501 on my first semester and learned a lot about basic common knowledge. However, too much homework in each week and only counts for 15%. HW are graded by students. Most of peer reviews are reasonable but a few of reviews are irresponsible. All exams are related to knowledge instead of coding.

Background - Limited Coding/Advanced Math/Statistical Background. Took several R-intro courses and the Linear Algebra EdX course prior to enrollment.

This course pulls no punches and throws you right into the fire from the onset. The first 3-4 weeks were very overwhelming (with my limited background) as I had to research the mathematical underpinnings of each model, the processes of each model, and the coding required to do each homework. I might have spent around 40 hours/week for the 1st month, pulling at my hairs and wondering what I got myself into.

That said, the structure of the course became quite easy to follow, especially once I got ahead and began taking notes as soon as weekly lectures opened up. The homeworks are designed to reinforce concepts and the meat of the course is in the lectures and its different model explanations.

If you can gather a strong understanding of the different models, you will succeed in this course. Statquest and other Youtube videos and resources online were great complements to the lectures and as the weeks went by, I was spending a lot less time overall on the course, albeit with a greater understanding of what I was learning.

Homeworks are fair and can teach you a lot about the models and R. Lectures are not overwhelming, though some modules can require multiple viewings. Exams are fair, though the wording of some questions can make you wince.

This course re-taught me how to learn in an educational setting. It also reinforced concepts on how to teach myself and research topics on my own. I, now, feel comfortable in my knowledge of the basics of most models and their purposes and have chosen subsequent courses to learn more about the models I want to study.

For everything up until the 1st Exam, the difficulty rating would have been a 5, and I was spending around 40+ hours a week on this course. After the 1st Exam, I would say the difficulty was around a 2-3 and I was spending closer to 15-20 hours on the course.

Conclusion - A very manageable course, even for people with minimal statistical/math/coding knowledge - Do not let the first couple weeks intimidate you and do not fall behind. Make sure you understand the models fully and you will succeed.

Background: experience with R from undergrad and knowledge of basic ML concepts/terms (which I think those 2 have benefited me really really greatly)

After midterm 1, I decided to read through the syllabus again to get a good understanding of what this course is meant to teach you, so I started to invest less time into going into every detail or formula mentioned in the lectures. Overall, I now feel fairly confident with my analytical instincts, for example, knowing what models and techniques to use when presented with a specific problem and dataset. The course project and case studies really drive home all the materials that I have learned. They were also my favorite parts of the course.

One regret that I had while taking this course was not paying attention to my stats class in undergrad :-)

Awesome class and a great intro to the program. Very light coverage of a lot of concepts, which helps to get some basic vocab in place to avoid diving into the harder classes in the program and getting lost. This is a class where I wish I could keep the lecture videos to refer back to throughout my career.

I wouldn't have taken this course if it weren't mandatory. I learnt a bunch of terms, thats about it. Wish they had replaced this course with a graduate level math/stats course laying a foundation for other courses in the program.

I don't get all the hype about this class. It was a good class for the students to learn new concepts about data science, but it was too fast-paced. I couldn't keep up with it (Yes, it's probably because it has to fit the summer semester schedule). Also, the first week R assignment was too difficult to pick up for those with little experience with R (It gets more accessible, or you just get used to it. I guess that's the intention of it being the first assignment). It can be challenging for those whose mother tongue isn't English to understand the lectures. It might be to do with many unfamiliar terms from math and statistics, but I had to replay Professor Joel Sokol's video many times to understand his way of explaining it. There are three exams, mid term1/2 and the final term. The mid-term one was the most difficult, and it took a lot of time to answer several questions. It's better to focus on the lecture videos, and it is not so necessary to do additional research on every lecture if your primary goal is to get a good grade. Overall, I am happy that I learned new analytic models and refreshed my memory on statistics and math. Still, the struggle to learn R and too many topics that I had to learn during the semester was brutal.

I really enjoyed this class. I think that the idea is really great and while I was an OMSCS student before, I am strongly considering OMSA now because of this class.

The issues with this class come from how it is structured. That is not to say that it's poorly designed but a few small tweaks could make this class better.

The first thing is there is a homework assignment, every, single, week. The only time we didn't have one was during our spring break. Every test, project, midterm, there was a homework assignment due that same week. The problem with this is that the homework is 15% of your grade. When you factor in the 3 quizzes (25%) this is almost laughable.

The other issue I have is that the homework isn't referenced on any quiz. You could legit never do a homework assignment for this class and still earn a B in this course if you understand the content. To me that's just poor design. The other issue I have with the homework is the fact that you have to attend office hours to figure out what to do (an hour each twice a week) and for the first quiz I spent SO much time on the homework because I thought it would reflect on the midterm, after I realized that it didn't and the main focus should have been the 20 min lecture videos, the class got easier.

TLDR, Don't stress on the homework. Focus on the content more than it. Even do outside reading/understanding to help your understanding but skip office hours and homework.

If you are an OMSA student, take this as your first course. You will learn all the concepts (at a high level) that will prepare you to survive in this program.

I think this is one of those courses that people review highly because they're enjoying their undeserved A.

What's wrong with this course is how they evaluate you:

• The exams are horrible and poorly worded. They don't test your knowledge but your ability to decipher questions written by someone who does not have a good grasp of English. Honestly, some questions felt like rolling the dice because I had no idea what they were asking, or I couldn't understand what they meant in the multiple-choice answers. This is someone who got As on all of the exams, so it's not just me being salty.
• Another reason this course sucks is that the assignments don't force you to learn. The TAs pretty much give you the answers during office hours. You can get by in this course by doing almost zero work and copying what the TAs did.

The good:

• The lectures are of high quality, but they don't redeem the course.

This was my first course in OMSA, which I'm doing to pivot my career from mechanical engineering to analytics. I knew nothing about analytics going in, and now I feel like I have the foundational knowledge I need to take the next steps.

I found this course to be reasonably challenging. I ended up getting an A and I worked hard for it. I'm working full time and found the balance to be challenging. To do well, focus most of your energy on truly understanding the lecture content. I was organized and made detailed outlines of each lecture which helped me prepare for the exams. I didn't hesitate to ask questions on Piazza if I was stuck on something.

The homework is an opportunity to apply what you've learned using code. Many other reviewers complained about how the homework takes a lot of time but isn't worth very much for your grade. To that I say: if the instructors wanted us to prioritize the homework then it would count for more of our grade, but they don't so it doesn't. Focus on deeply understanding the concepts from the lectures, which homework can help supplement, but don't be a perfectionist on the homework unless you have the time and energy to do so. I had 2 zeros (dropped), a 50, a few 75's, and I still got an A because my exams and course project were all solid. The course project (8%), which took me as long as 1-2 homework assignments, is worth half the homework (15%), and the 3 exams (75%) are worth as much as 5x the homework.

Many other reviewers also complain about the home peer reviews lacking quality feedback. I agree, but I also accept that as part of the trade-off for such an economical degree. I'd like more feedback but am unwilling to pay for a more expensive program that has more resources.

This ended up being one of the most useful classes for my day to day job. I wouldn't say it was the most fun, but it was the most useful course I took. You'll do a shallow dive into many analytics models, but one of the main purposes is to learn enough to know when to use each type of model. The homeworks are nearly relentless, as each week you'll dive into something new. Most is in R, but a couple homeworks can be done in python, so if you're new to either it may be an opportunity to learn. Exams are very fair, but are proctored. As some mentioned the assignments are peer graded, but you will get out of it what you put into it. By the end you'll realize you learned a good bit by digging, and not worry about how someone will grade you as much.

This was one of the worst classes of my entire OMSCS experience (out of 10 total courses), which is a shame because the material has the potential to be interesting. The class is clumsily delivered using kaltura, aka a video player straight from the late 90s, and makes for an atrocious user experience (why can't you use ed lessons like most of the other omscs courses?).

The homeworks are graded by your peers. This is fine in theory, but in reality it means no one is actually evaluating your answers (on a homework I got every question wrong on I got 100%, and on another I got everything right i got 90%, etc).

The exams are the worst part of the course. They are essentially trivia recall, which is a peculiar way to test analytics understanding. If you happened to write the right formula or sentence from the lectures onto your "cheat sheet" you'll be fine, but conceptually understanding the vast amount of material won't get you very far. The exam questions didn't appear to be written by a native english speaker and much of the exam time is spent trying to decode what the ambiguously-worded question is actually asking for. This is an absolutely terrible way to test knowledge and is not becoming of a "graduate" level course.

I'd pass on this one if you can

This was one of my favorite classes in the program, especially for someone with little to no machine learning background. It does a great job of introducing a lot of Analytics concepts in a way that makes you feel like you are learning a lot without being totally overwhelming. Professor Sokol gives lectures that seem to be as concise and clear as possible and do a great job of communicating information.

R was really the only programming language used, and maybe Excel for a homework or two, nothing too crazy.

Interesting introductory class to the program. Depending on your background, this class will either make a good jumping off point into data science or a good refresher course. I highly recommend supplementing the lectures with the relevant chapters in "An Introduction to Statistical Learning", this is a free textbook easily found online if you are unaware. The homeworks are peer reviewed which was never an issue for me but don't expect to receive any feedback of substance from the reviews. The exams have some trick questions that require careful reading but if you pay attention to the lectures you should do fine. Overall a good course for what it is.

This was the first course I took, and provided a really solid overview of analytics modelling and the degree. First two weeks felt challenging given my limited experience using R (I had done the recommended pre-requisite edX course and nothing else). However, I quickly learned the TAs pretty much showed the code you needed to write in Monday office hours - no need to sweat then, when I could sit tight, copy their work and then do a writeup.

I really enjoyed the content, exams weren't too easy or too difficult, and weeks 11-13 homework that let us apply knowledge were fun, and so was the project. The only thing that was slightly upsetting was peer grading - your grade could vary depending on who happened to do the grading, and more often than not I wouldn't get any written feedback.

Overall, this is a good course.

If you want the long version of my review you can read it here: https://medium.com/@jonathanbechtel/course-review-isye-6501-introduction-to-analytics-modeling-79b2bd0376b3

Here is a concise overview of what's said there:

• Class is a true survey course that covers a wide range of topics with modest detail
• Emphasis is on intuition for how/when to apply different methods moreso than mathematically deriving them
• Most of your grade comes from exams, and exams come from lectures, so make sure to watch them carefully
• Use homework to understand lectures and to make sure you get the ideas behind different concepts, but do not obsess over the last detail with them. They are worth 25% of your grade total, and you have 14 of them - so each one is a small portion of your grade
• HW is peer graded, so expect your grades to be a little inconsistent
• Workload is a little front loaded since there is no introduction to R. The last three weeks are built around case studies and require no coding. Last 3 weeks only require a couple hours / week
• If you are new to coding the HW can be time consuming. The coding itself is not so complicated, but be prepared to work through R the first few weeks if you haven't used it before

Overall, I liked the class and would recommend it as a first class.

I really liked Prof. Sokol’s delivery of the material – and I really wanted to like this course. However, as a former teacher (and education researcher), I feel there are some aspects of this course that need to be revisited simply in the spirit of good teaching. To keep this short, I’ll just mention the two that irked me the most: the peer reviews and the exam format.

First, peer reviews. I think it’s a great opportunity for your peers to review your work (and vice versa), but peer reviews are neither helpful nor fair when they are based on the vaguest rubric I’ve ever seen. It just results in all HWs being regressed to the de facto mean of 90% -- even when the quality of the work is far above (or below) that level. Let’s be real: the purpose of peer reviews (as they currently exist in this class) is to get grading done more efficiently. If they were meant to be used as a learning tool, there would be a much more detailed rubric or alternative process for grading HW assignments.

Second, exam questions. I was a little taken aback that a graduate-level course (however introductory) had such simple recall questions for 80+% of the exams. In addition to that (as the previous reviewer mentioned) the wording was so horribly contrived on some of these questions that answering them felt more like a logic game than an analytics modeling exam. Difficulty aside, I’ve seen so many MOOCs have so much more engaging exams.

Designing an analytics survey course with the right balance of concepts, tools, etc. is probably a big challenge, so I applaud Prof. Sokol for that. The TAs clearly work hard, too, answering questions and doing HW regrades. You would just think that in the introductory OMSA class, a little more attention would be paid to pedagogy.

(If it matters to whoever is reading this, I got an 'A' in the class.)

There are aspects of this course which I strongly liked, while others I strongly disliked.

First of all, while R and Calculus are definitely used, Statistics should have also been mentioned as a prerequisite. As someone with a strong calculus background, weak stats and 0 R, I found myself struggling the most with stats concepts. R was a struggle during the first two assignments, but it's a very easy coding language and if you have an extra hour to spare every week you can easily catch on.

The homework assignments are enjoyable, but very inconsistent in terms of workload (one week it will take you about an hour, another week it can take up to eight), and peer reviews need to be better monitored. There doesn't seem to be any obligation for a peer to review you fairly as long as they submit a review.

Additionally, the exam questions need to be beta-read by speakers from different countries! I'm not a native speaker but I like to believe I'm fluent, however I struggled a lot with language ambiguity in the exams. ESPECIALLY since this is a program based on mathematical concepts, I would assume language clarity to be vital. If the answer to a question is unique and absolute, then the question should reflect that (for example: you wouldn't answer 4 to a question that asked what is x + y if x is 2 and y may or may not be somewhere around the same value). Plus, there are also inconsistencies between versions, this could be the difference between an A and a B. As a graduate level course, I would expect there to be more care put into making the exams as fair as possible.

While I would have liked more interaction from the professor, the lectures were all enjoyable and Dr. Joel presents concepts in a very easily digestible way. I appreciate his enthusiasm because it made re-watching lectures painless. Most of the TAs were very helpful, but some could definitely "dumb" themselves down while explaining concepts in an introductory course. I liked the content and I felt like I learned a lot. I definitely would have enjoyed the course heaps more if the assignments and exams were reconsidered.

This is the definition of a survey course, and the structure is strange in that the homeworks are very focused on implementation of models in R while the exams and project are very focused on model assumptions and case studies. There is often very little overlap between the two, making it feel like two separate courses sometimes.

You don't realistically need to know R before starting; you will not do much beyond a few for loops and calling methods from libraries. If you had no coding experience in any language, this would probably make for a hard time.

The listed academic prereqs are a little exaggerated--you do not need to know linear algebra or calculus to get an A in this course. One college statistics course is probably sufficient.

The lecture videos are great and contain everything you need to succeed on the exams--they're nice and short and can be rewatched at 1.5x speed before exams as a review without taking an inordinate amount of time.

My one gripe... Don't make me handwrite a carefully designed cheat sheet. Please. I just went without on the second two exams and was fine.

This course has a few major flaws:

Inconsistent difficulty gated by prerequisite knowledge: If you know how to code already, homeworks take an hour. If you took basic statistics a year ago, you are fine. If not, study up. The exams questions are either too easy or purposefully worded to trick you. Spend most of your time on exam prep.

TAs: I had great TAs and terrible TAs in this course. Some of them were fantastic, others ….. didn’t even know what dplyr or ggplot were despite this being a primarily R and data science oriented course. ??? That’s like not even knowing what pandas is in python, and these people are grad students? What a joke.

Prof: Did we have a professor? He never showed up. Overall his lectures were fine and funny, but I could have gotten the same out of reading introduction to statistical learning. It was embarrassing that in my analytics masters my first business class had more professor interaction. Not a good look, but maybe that's just the prof.

Overall I did fine (an A pre-final) in this course and learned a bit, but don’t hold your breath thinking it will be anything special. Some courses are just meh… I guess this doesn't change in graduate school

This was one of the lighter courses I have taken where homework did not take me a super long time to complete. This class focuses on breadth rather than depth so you get an inch deep understanding of every topic in analytics. I feel like this course could be useful for interviews where you get asked a bunch of "what does this do, what is this.." type of questions.

I did not like the peer reviewing system. It just seemed like a way to not have TAs do any work which is why they seem to have a ton of time answering piazza questions immediately.

The exams were tricky and I did pretty mediocre on them. The wording can throw you off a bit and it's hard to know how to study for them. You really have to understand your concepts and how they apply to other situations.

I would say an improvement to the course would be at least give some sort of introduction to programming in R as you get thrown into the deep end and they give you very little resources to figure it out. The R documentation is very poor (I've been spoiled by Python docs) and sometimes you are left puzzled on how to do stuff. You will get out of this class with R knowledge so if that is what you want, this will do it.

Thank you Joel Sokol for blessing us with this gem of a course!

Genuinely the best course I have taken, tied with Andrew Ng's Coursera ML class.

This course covers a nice overview of the more common analytics models in a way that is easy to understand without too much of a math background. I liked that he focused on many different types of modeling approaches, but really drilled into important fundamentals that are relevant across many types of modeling.

This course lays the groundwork for many other courses in the OMSA program, such as Regression, Time Series, Deterministic Opt, Simulation, Stochastic Processes, Deep Learning, and more. I have yet to take these courses in the OMSA program (I have taken undergrad versions of some of these) so I am looking forward to getting more into it.

The homeworks were time-consuming but important for practice applying the techniques used in the course and to gain some more exposure to R.

If you take this course and actually understand what is going on, I think you can transition from an Analyst to a Data Scientist. For me, I am in that process now - looking for ways to start applying these techniques to my job.

Basic overview of analytical methods, decent survey course to learn a bit about many approaches. Low workload, peer review system made getting feedback a little difficult and the requirement to work in R was a bit annoying.

It is true that the course is a great introduction to the entire program. Even though the course touches upon a bunch of concepts, there is enough detail to keep one interested. There are weekly assignments, which can get annoying, but they keep you on track and help with understanding the content on a deeper level. I was not a big fan of the midterm/final, which was multiple choice. Personally, I would have liked to see some R coding in the midterm/final since the assignments relied heavily on it.

You'll hit the ground running with R immediately. This is a very useful and interesting course but have some preparation in R before you start.

I love the material for this course, it covers many things analytics need. I wish to learn little deeper for each topic, but this is just an intro. (I do recommend to do some outside reading for each topic, so you could have a better understanding.)

There is 1 hw per week, using R. Some very easy some takes time.

The only downside is the terrible wording in the exams that led to ambiguity. I didn't do so well in exam 1 due to not familiar with the wording and misunderstand quite few questions. Be careful on that.

I took this class with CS 6400 in my first semester of OMSCS since I did my undergrad at GT in ISYE. This course was easier than most classes from my undergrad and didn't give me the same "Georgia Tech" feel -- might be because the teaching staff was extremely outgoing and Piazza was really active, but I don't think a graduate course should do as much hand-holding as this one.

Overall the class was what it said it would be; inch-deep exploration into a wide range of ML modeling techniques. Homeworks were weekly and peer-graded, which was pretty cringe-y (lots of overnight ML experts leaving notes like "the videos didn't mention eliminating variables in regression models" or "your randomly-partitioned data didn't result in the exact same accuracy as the professor's solution" ...). The default peer grade per the prof should have been a 90 unless you were significantly off the mark, but the class average hovered around 70-82 most weeks. If you take this class, submit your code and analysis as a R markdown PDF that clearly steps through your process or risk getting a bad grade from lack of readability.

Exams were multiple choice and fairly straightforward, but there were occasionally "choose the best of these options" questions that could throw you for a loop. Didn't watch office hours but apparently they gave away the code needed to complete the homework assignments so no worries if you're new to R. Actually, a lot of people on Piazza/Slack mentioned never programming before this class -- though I'm in OMSCS, I would have assumed that OMSA students shouldn't be entirely new to programming or at least expected to be able to learn on their feet enough to not need most of the homework solution given to them each week after already being told which libraries/functions to use. Embarrassing to see as a Tech alumna.

I really like this course as it touched on many topics. IT gave me good overview what I would like to take next. It time intensive for me because I have never used R but know some Python. R has so many libraries so I was a bit overwhelmed for the first two assignments. Very good introduction to OMSA.

Overall it was a very good course. It's best to take this one after CSE-6040 if you don't have a lot of programming experience, as you are expected to learn R without any direct instruction, which actually is not very difficult if you have experience in another high-level language.

The programming exercises are rewarding and meaningful. The case studies are just ok and should have been spread throughout the program and the final project is somewhat ambiguous. You are at the mercy of peer graders for these. The good news is that these grades aren't worth a huge amount, so even if you don't get 100s, you are good if you like taking exams. I got mostly 100s and a few 90s and they are generous with averages if you have a jerk who gives you a 75 for a correct solution.

The exams can be tough. It's not really the material. Instead, it is the way the questions are asked. You have to get used to a Joel Sokol exam, as the kinds of questions you are asked are strange. The good news is that most of them are multiple choice or binary choice, so it's REALLY hard to fail the exams, but also really hard to get a 90+ on them either.

They also don't share grading cutoffs, so if you're left on like an 89, you have to wait until grades are released. I find this lack of transparency frustrating, but overall a good course.

As many have already said, this is a great high level overview of analytics. This is my first class in this field since taking my undergrad in a related major (Operations Management). It's clear that the field has advanced quite a bit in ~15 years, and I'm glad I took this class to update my skills.

Difficulty: the material is challenging but not unreasonable. I consider it a similar difficulty level to some of my harder senior undergraduate classes. Several of the weeks were very straightforward (Basic Regression), and some were quite challenging (Principal Component Analysis)

Learning and Delivery: excellent delivery of the lectures, they are ~40 minutes per week, but the amount of material is condensed when compared to live lectures, so it's worthwhile reviewing the material a couple times. The office hours were mostly great, and sometimes fairly lousy (all depending on the teaching style of the TA), but they are the primary source of information on how to complete the homeworks. I learned a ton of great info on the homework assignments, this is where I found nearly all of the practical knowledge in this course. The exams were reasonable, I like the format, but others may not like the many matching or multiple choice questions. Open ended would likely be more practical, but there are obvious cost savings in the format they chose (and this program is all about bang for the buck).

Time Commitment: Many weeks I'd only need one evening to watch the lectures/take notes and one to two evenings to do the homework assignments. I'd normally watch the Thursday office hours (recorded), do 80% of the assignment, then finish it off from any material gained from the Monday office hours. I would've loved to devote more time to the material because I find this stuff quite interesting, but I just wasn't able to due to work and parenting commitments.

Pre-requisite Knowledge: make sure you know how to program well and have the basics down for R before even starting this class. Beyond this, make sure in the first week you sort out all the IT knowledge right away. This includes R Studio install, R install, learn the proctoring software, knitting R Markdown files to PDF (harder than it sounds). There were other undergrad classes as pre-reqs that I'd recommend like Statistics, Calculus, Linear Algebra. If you have 2/3 of these then you should be good.

Final Thoughts: I like the peer grading, it gives me a chance to view other people's work and see interesting techniques, so be it if some don't even bother to read your assignment submission. I also don't mind the online aspect of it, as the TAs are helpful and it allows me flexibility of timing to complete the work. This could be essential for those with family/work commitments.

Sure there isn't a lot of depth to this class (just breadth), but my role it work is more high level and not hardcore data mining, so just knowing the basics and the terminology is half the battle. From here I can more easily work with the IT groups at my company that will actually do the hardcore data work. If you work with IT/Data people, this is probably the only course you'll need; however, if you want to actually do data science work, you'll need to complete several more classes to be effective.

Overall this was a great class, and the professor's dad jokes were on point.

I loved this class. I came in with a tiny bit of R experience, but a decent programming background otherwise, and did not find the coding to be that challenging (I imagine it would be a lot harder with no programming experience).

This class is well-designed and the Professor does a great job laying out the material, explaining the math behind it (as needed), and providing applications for each of the models discussed. I also liked that the homeworks were directly related to the lecture content, although they did require outside research. This matches my experience in work and in other educational programs though - I've never taken a class where you could be 100% successful solely by engaging with the class materials and without bringing in outside resources.

I did not love the peer grading, mostly because it occasionally felt like people weren't being thorough and the comments I got back were almost never helpful (lots of "good jobs!"), but I enjoyed the process myself.

People talk a lot about the "trickiness" of the exams. I strongly disagree. I thought they were suitably challenging, but if you watch all the lecture videos, do all the homeworks, and fill in any gaps in your understanding with other resources, you will succeed. I did not find a single "trick question" on any of the exams - understanding nuance is not the same as a professor trying to trick you.

IYSE 6501 Review - Sp 2021

This is an absolutely fantastic course. If you have to take a single course in the entire AI/Analytics/ML areas at Ga Tech - this is the course. It covers everything and gives you a framework for further exploration. It is a little slanted towards a Statistics practice (than say AI/Comp. Science), but there is overlap (I have taken AI, and some topics are in both classes). The "unofficial textbook" is Intro to Statistical Learning. But you do not need it. I read the first few chapters and the chapter on SVMs, and the material in the text is much more advanced than the course expectations. I stopped reading it after the second week (but it is a summer goal to read the entire text. Oh, the ambitions of a simple man!). Stick with the videos, Piazza, working on homework, and the "knowledge checks," which are little quizzes after the week's videos.

The Big Picture This course is exceptionally well organized - from the videos to the homework and even the instructors on Piazza. I will cover each one in a little more detail, but here is a summary of what to expect.

This course is harder than it may seem at first, and it can sneak up on you. The pace starts out slow, and it is easy to fall behind. By the time you have to take the first midterm, you realize there's a ton of material to prepare for the first midterm. The second midterm is less than a month after the first, and the work deliverables keep comping. The Project is due right after/around the second midterm, and you still have homework due. It can be a lot to balance. My advice is to keep on top of the assignments and videos and try not to fall behind - catching up can be a little challenging. Easier said than done.

The curve seems pretty high, but they do not publish it. My guess is that most people end up with a B, but you can get an A if you are really good. The midterms carry most of the points (about 75%), with the remaining 25% divided between the Project (7%) and the homework (15%).

Midterms Midterms are long and tricky. They are fair, but some of the questions are wordy and nuanced with lots of build-up with background context. You have to read them carefully because they want you to choose between subtle features in various models, etc. They allow you to have a cheat sheet. For the most part, I rarely used it - but YMMV.

The best way to study - I think - is to watch the videos. The videos are excellent. A previous reviewer complained that Dr. Sokol had a monotonous delivery, but I disagree. I actually liked his presentation style. It was very clear, well thought out, and Dr. Sokol has a fine delivery. I watched them about three times total and learned something each time. But the enjoyability of the videos is clearly a subjective assessment.

The final is cumulative - 65 questions and 3 hours. It really is cumulative, and there is a lot of material to choose from. Towards the end of the class, there is a new question type that occurs. The new question types ask you to apply the models you have learned during the course - why did this one work, and why will this maybe not work. You have to know your stuff.

Homeworks Homeworks are R-programs that cover the topic of the week. Some of the problems were actually pretty difficult, but the grading is very generous. Also, the TAs hold a session where they walk you thru each homework and help you set it up. The point is to get a hands-on feel to the theory. Assignments are graded by your peers, and as long as you give it your best shot, the grading is generous. The Project is like a more extensive homework assignment, and they ask you to apply analytics to a public problem you can choose from.

Piazza The instructors and TAs are sensational! I did not realize this until the latter half of the course, but you can ask any question you have on the material on Piazza, and the instructors WILL answer all your questions - every single time. And they really do a great job. There were many great conversations on Piazza with great student and instructor involvement. Piazza was a great resource, especially for exam prep and The Project.

Summary This course lays the foundation for a career in Data Science. It covers a lot of material in sufficient enough detail to know how to get started on a problem. This was my first class from the IYSE side of the Ga Tech OMSC program, and I must say - if other courses in IYSE are similar in quality, I regret not applying for the Analytics Master instead of the Computer Science ML masters. (Both are good, I just really learned a lot in this course).

Just finished 6501 as my first of two courses in the program. My background in data science made this course easier than it would be for many others - I would estimate that half or more of the material was review for me, although a lot of the material comprised of concepts that I learned about years ago and was quite rusty at. If the material will all be new to you or you do not have some programming/R experience, this course will require a few more hours per week to be successful.

What I like about this course:

• Breadth of material
• Focus on applications
• The homework assignments forced me to re-learn R, which I had not used in a long time
• Professor Sokol is a great lecturer
• Lots of great questions and discussions in Piazza and Slack. I never had trouble getting help.

What I did not like about this course:

• Starting off the material with SVMs...SVMs are one of the more complicated statistical models, and IMO should be taught after regression and other more foundational concepts. Part of me thinks the goal here is to weed people out of the course quickly. This might be controversial, but SVMs are also used infrequently in practice (although they are still useful to learn) and going into depth on regression or decision trees/random forests will be much more helpful as prep for working in industry.
• Not enough depth on the algorithms presented. I know the aim of this course is breadth and not depth, but some of the concepts felt too rushed. I would pare back the algorithms by 10-20% in order to create more time for doing a deep dive on even a couple specific algorithms.

The exams are challenging but fair. The homeworks are not bad - like another reviewer suggested, I realized during the later half of the course that starting the homework late was very useful because the office hours and people's questions that had been answered in Piazza made completing the homeworks far easier.

Good course:)

To succeed at this course I'd suggest the following:

1. Watch all the lectures and create flashcards out of them. I've used the 'ankiweb' service for flashcards. I've watched lectures once, then created flashcards out of them, then I've watched the lectures again to make sure I haven't missed anything >> Saved flashcards in anki.com. That's it. I never watched any lectures after that. I've only repeated flashcards on anki.com and that all.
2. To succeed at Homeworks, I suggest that you start them ONLY AFTER watching Monday office hours. They give you 3/4 of the code and you just need to complete the rest.
3. I recommend being patient at quizzes. You have 1.5 hours per two first midterms and 3 hours for the final one: make sure that you use each minute and only submit the test when it is 3-5 minutes left. If you have done your homework, watched lectures, and repeated flashcards created from them, then you should be fine. I believe that your grade would be between 80-90% using this strategy which is quite good.

Lastly, I'd highly recommend doing the following to prepare for the course:

1. Do some R programming courses: For example, in DataQuest (I highly recommend 'Data Analyst in R' path) or DataCamp .
2. Do prof. Goldsman's course on Statistics and Probability on Edx, if you haven't studied those in undergrad.
3. I also recommend at least some knowledge of Calculus and Linear Algebra. I recommend the 'mathtutordvd.com' site to refresh your knowledge in that area.

That's it. Overall quite a nice course. It helped me get two job offers recently. So, I suggest that you listen carefully to what is said in lectures since the same type of knowledge is often required in the interview.

I'll start with the negative comments just to get them out of the way...

1. The course is very fast-paced. I feel like there were times when I didn't really learn much just because I was having to move so fast.

2. I don't like the grading system. A 90% is considered correct but a 100% is going above and beyond. Kind of strange. But the overall contribution of homework assignments to your final grade is pretty small. So it probably doesn't make that big of a difference (but I haven't run the numbers).

3. The homework assignments are a bit vague at times.

That's all the bad!

The content was very interesting. I don't have an analytics background whatsoever, so this was a great course to hit some hot topics and scratch the surface of the field. The lectures were very good in my opinion. Don't be fooled by how short they are...I spent a lot of time pausing the videos to write notes because it's so fast paced (you don't have a chance to write while the professor writes because it all just pops into the screen). The lectures weren't very useful for the homework assignments, but extremely relevant to the exams since there is no coding on exams.

I think the TAs did an amazing job on Piazza. It'd be interesting to know how many responses they gave over the course of the semester! Office hours were pretty good. There isn't a lot of question and answer time though, mostly just TAs discussing the assignments.

My last blurb here is regarding people complaining about the number of people in the course, peer grading, lack of instructor interaction, etc. Yes, these things are unfortunate. But we're getting a degree for $1000 per credit hour less than most universities. Corners have to be cut to knock the price of an online graduate degree down to this price.

I hear Prof. Sokol is so seasoned that he initializes model parameters at the global optimum Ö

Things I really appreciate:

1. Prof. Sokol’s love for the subject is really obvious & almost makes up for the fact that ISYE6501 uses R. Also, his vast wealth of real-life analytics experience across multiple industries really helps ground our understanding, broaden our perspectives on various edge cases & drive home the point that modeling is an art. Excellent introductory course to analytics.
2. The format of weekly discussion questions on Piazza is really helpful for aggregating thoughtful comments on each week’s content, & separating it from the noise. Super helpful during exam revision.
3. The large cohort size means every cohort will have several outstanding students who contribute invaluably to Piazza discussions. Also, with many students learning the same things & looking up the same things on the internet, there are many Piazza posts sharing articles/papers from Medium/other schools that explain concepts very well, which really helps to clarify & broaden understanding.
4. The “open-ended” analytics questions towards the end of the course, and the course project, really help us to think through how various models can be combined to answer real-life questions. This really brings the entire syllabus together.

Things I don’t like:

1. R. Horrible documentation, multiple libraries doing the same thing, weird lack of interoperability between objects (vectors, tables, matrices, dataframes), plus an annoying tendency of statistical libraries to “print” rather than “return”, meaning you gotta copy & paste the values again manually. If you don't know this language, you'll spend hours more each week just sifting through the little scraps of documentation available. Apparently I've been spoiled by Python documentation.
2. Grading system: Too many homeworks where people just run models mindlessly & regurgitate the results. Rubrics should allow greater penalties for insufficient qualitative analysis & analytical thinking. Homework weightage should be way higher to incentivize more effort, (CSE6040’s homework is 50%). We should also be allowed to penalize lousy code. I’ve seen cases where the code stretched 4 pages long just cleaning data.
3. There’s a really weird lack of practice problems come exams; mid-term 2 didn’t even have any. Letting students practise past exams (as does CSE6040) really makes a difference not just to scores but to learning. Doesn’t help that the Prof. loves setting trick questions, and that exams are 75% weightage.
4. Inflexible: CS7646 has spoilt me in that all projects are posted upfront & you can tackle them in your own time. Here, you’ve to follow a strict schedule, which isn’t the best when layering on the vicissitudes of work life.
5. Peer grading: All homeworks are peer graded by default; TAs only step in when a peer did not grade you. I’ve made several mistakes in homeworks that my peers didn’t even notice. And almost always, comments for my assignments were just one-liners, if present at all.
6. By the time we get to the “open-ended” analytics assignments, which is the main meat of the course, a lot of students are too preoccupied with exams (or they just don’t care), which means you don’t always get to see good examples.

NEUTRAL word of caution: mathematical notation is notoriously inconsistent & it’s no different here. Lecture formulas can differ from other papers/disciplines/libraries, and this starts right from Week 1 with SVMs (λ vs C). Some inattentive students only realize these inconsistencies when they’re >10 weeks in. But it’s not really the fault of ISYE6501 – just how it is.

I took this class knowing the basics of R syntax. Having the title of "INTRODUCTION to Analytics Modeling," I mistakenly thought that we would have had some sort of INTRODUCTION to modeilng. The first week's assignment was to do linear regression to predict stuff in R. Dr. Sokol's course videos really do explain the concepts very well in extremely short chunks. However, his philosophy seems to line up with the traditional drownproofing method.

Optional background assignments like in CSE6040 would have been helpful. There was zero background resources (some of which are licensed by Georgia Tech) linked to from the class. When this was suggested, the TAs got really defensive and went to the typical "you're a graduate student, figure it out." This is not my first graduate degree or course, so this is really patronizing and unhelpful. My biggest issue was knowing what the output meant. R documentation is horrendous and googling "how to do X in R" is frequently unhelpful. Apparently I've been spoiled by python documentation.

I feel like if the first two weeks had been dedicated to regression, how important it was, and how it works in R, and what the output meant, I would have really benefited.

Other than an occasional Piazza post, there was no interaction between students and Dr. Sokol. The office hours were 100% run by the TAs and Dr. Sokol never attended. Out of 4 classes I've taken, this is the only course that I have had zero interaction with the instructor of record. The TAs were vague and often didn't answer questions because we needed to figure it out. Consistency of TA holding office hours would have been nice. Some did a really great job, and others were terrible. Since there are so many TA's, having a head TA in charge of office hours would provide improved quality. Having Dr. Sokol attend occasionally even at the first office hours could have provided some motivation and encouragement.

The exams were very tricky. Not the content, but the wording was designed to trick you. The homework was OK rigor-wise, but the peer review rubrics are ridiculous since a correct homework answer is only worth 90% - all homeworks are peer reviewed by 3 classmates who are also clueless. Grading by them was fair but in order to get 100%, you had to go over and above the requirements, which was determined by the peer graders. If you get a correct assignment, then you should get full credit. if you want to make it harder and then give 100% for fully completing, then 90% if you didn't do the last question, and that would be a clear way of doing this.

I fully understand that this is a survey course, but also this is intended to be an introduction to the program...for many the first course. Luckily I took this as the last of the core requirements so I was not scared away by this course. I ended up getting an A in the course but felt that I could have gotten more out of it had there been additional resources included.

The pace in this class is relentless. It's so fast you have no time to dive into anything that you find interesting.

The week-to-week lectures are all over the place. One week you're learning clustering, then basic data preparation, then change detection, then time series, then advanced data prep. There doesn't seem to be an order what so ever. I honestly think two weeks in a row would have been nice to spend on linear regression given how important it is. Instead Dr. Sokol is attempting to cram as much material into an "intro" course as possible.

Speaking of "intro" course, the lectures touch upon the material with "intro" depth, but the exams test the students as if they have been working in the field for years. Make sure you know EVERY single detail about every single lecture because you will be tested on it.

The exams are extremely tricky. Don't be fooled and believe since you have a good grasp on the material that you will do good on the exams. The wording is meant to trick you.

Overall I highly regret taking this class. I could have learned way more by buying an intro to statistical learning textbook and taking my time. There is no time to learn anything. You need to keep chugging along, submitting hours worth of homework every week. The homeworks are peer reviewed by 3 other random students. Make sure you attend office hours because if you don't implement the homework solution exactly as they describe in the office hours, the students will start you out at a 75 and it hurts. In the beginning of the semester one of the TAs even tells the class that the baseline for a correct and complete homework should be a 90. That's insane. A correct and complete homework should be a 100. I essentially needed to submit research project to get a 100 on the homework.

I wouldn't recommend this class. Pick up a statistical learning textbook and save yourself 900 bucks. They make this class hard and dull for no reason what so ever.

Pluses:

1. Good short loaded with info video lectures

Minuses:

1. Horrible TAs (majority of them, with some exceptions). Horrible TA sessions. Some other courses have more than 7 TA sessions per week. This class has only two. Students still have a lot of questions when the OH is over and are encouraged to post them on Piazza.
2. Piazza is overwhelming. With more than 1000 people in the class I DO NOT have time to read all of the messages, regardless how good they are.
3. Homework is very disconnected from lectures.
4. Prof. Sokol never showed up for any of the OHs.

I learned so much! This class was a perfect intro and made me want to learn more. You will have a serious leg up if you're previously exposed to R. There is also some Python.

This a great introductory course that provides a comprehensive overview of the landscape of analytical modeling techniques. The lectures are great and engaging! The weekly homework could be time-consuming depending on the topic so be sure to attend the weekly office hours to get help. I thought the exams were tricky due to the nature of multiple choice questions. I spent a lot of time on the homeworks since I'm new to R, and unfortunately the efforts on the homework didn't quite translate to the exams since they were primarily based on the lectures. Overall, this class was a lot of work and moved fairly quickly but extremely rewarding given the amount of material I learned and hands-on coding experience in R.

To give you my background, I had a BA in Econ & Math and have been working in analytics field for 2 years. I only have rusty intermediate level Python and SQL knowledge.

The course title should set your expectations for this course. This course gives you an introduction to many of the different analytical models that you will encounter in analytics. If you are trying to learn in depth about any of them, you are in a wrong course.

I thought the course did a very good job introducing these diverse topics. The lectures were mostly on point, though the professor looked a bit awkward on camera. Follow many of the details on the lecture carefully. The professor isn't as engaging on Piazza as some of the other professors, but he does give lengthy comments to different things every once in a while.

There are a total of 13 homeworks worth a total of 15% of your grade. The lowest 2 get dropped though. There was one homework that required you to use ARENA simulation, one that required PuLP (Python), 2 were just write ups, and the rest of them used R. The syllabus says even if you don't have R experience at all, if you are willing to learn it on the fly, you should be good to go. I didn't have any experience in R, but I thought it was manageable, especially because TAs give so much help in the coding aspect in the office hours. Some people were having a mental breakdown at the beginning of the course because they were expecting a lot easier coding, and although it will be challenging if you never used R, I don't think it was worth insulting the course staff and having a mental breakdown. If you follow the office hours, you should be passing. Make sure to download RStudio and get a bit familiar with it in advance though.

The homeworks are also peer graded on a scale of 50, 75, 90, and 100. I usually got 90s and 100s, so as long as you get most of the parts right, you should be good to go. I unfortunately didn't learn at all from the peer reviews, but I thought the numeric grades I got were pretty fair.

There is a course project worth 8% of your grade, and it is very similar to the last 2 homeworks (write-up about what analytical questions you are trying to solve and how you can go about to solve them). They are also graded in the same way as homeworks, but do try to go for a 100 on this if possible, since I think I saw many people on the borderline grades.

There are 2 midterms and 1 final, each worth 25%. You get 1 front-and-back cheat sheet for each midterms and 2 for the final. The exams don't test you on any math nor any coding, but you do have to know the theoretical aspects of different models very well. It is a bit tricky because a lot of questions make you choose more than 1 correct answer, but I thought they were mostly fair except for 1 or 2 questions that were a bit confusing in wording.

Overall comment: very good introduction course on many of the models we will encounter. Follow lectures carefully, and focus on learning concepts rather than memorizing equations. Get familiar with RStudio and at least the very basics of R, but TAs are very helpful in R coding for homeworks.

Big shout out to the TAs who were extremely helpful in office hours, Piazza, and Slack. Always seemed like someone was there to respond. Makes me feel more comfortable about the online nature of the program and its benefits -- 24/7 help and support!

The course itself was a great overview of many different analytical models. You should not expect to be an expert of any of them after taking the course -- my opinion is that this is just to whet your appetite and show you all of the things that are out there -- "you don't know what you don't know".

The videos are short and to the point with not a lot of fluff.

Homeworks are due every week but if you go to office hours they are fairly trivial.

I enjoyed doing and receiving peer reviews -- hope more classes have peer reviews!

This is a very broad course, very useful to get a grasp of different tools you can use for data analysis.

Positive>

• Great lectures explaining the how and why of different techniques. Not so much about the math under the hood.
• Interesting assignments mostly in R. These are useful to really understand the lectures.
• I liked the 3 exams format, I think these are good at testing the knowledge of the course. Additional information> For the first exam they give you some sample questions a couple weeks before, take advantage of that and do your best in this one.

Negative>

• Dissapointing final project. It has the purpose to let you solve a problem from a high level view, but with no implementation. Useful for the final exam and real life but isn't as exciting as the projects from other courses.
• The lectures and assignments are given week by week. There is no chance to advance and frontload the course.

A great introductory course that provides just enough depth to make you want to explore the topics in future courses.

Many of the complaints from students are typically from students who were not prepared or who don't understand the purpose of the course and the purpose of the peer reviews. It's all laid out nicely in the syllabus why peer reviews are done. Anyone who criticizes the course or professor for this has not understood the purpose.

The TAs were heavily engaged and give you just enough information to go on. If you actually go to the office hours, one could argue they actually give you too much and basically tell you how to do all the homeworks. People who complain that the TAs are impersonal are the ones who expect to be handed a gift-wrapped A+ in the course because they don't want to put any independent effort in.

I was very disappointed with the way this class is structured.

Video recordings: These recordings are about 30 to 40 minutes long on average. They are good, engaging videos with knowledge check questions, but not long enough and with not enough content to get you through the homework. This is why it takes you longer to finish an assignment due to outsourcing and researching, which is ok but having to do an assignment every week is a little too much for a graduate degree.

Peer Reviews: Each and every assignment is peered reviewed by an unknowledgeable and lost student just like you..very few students have actual experience and knowledge with these type of assignments...few to none, I should say. Keep that in mind when you read the comments not aligning with what you actually do/write/create in your assignment! Having that out of the way, just know that the homework percentage weight is not that much. However, it is very frustrating that students have this type of power reviewing assignments, because they can be rude and condescending when leaving comments even though they have no knowledge of what they are even writing. You will notice that some did not even read your assignment based on their comments, and on many occasions, if not all, they will just say "Good Job!" without good feedback. So do not expect to learn much from the peer review comments.

Grades: These are pretty straight forward and just keep in mind that the curve is not that significant and each exam is worth 25% of your grade. Make sure you study for the exam like you are in a Data Analytics interview with example questions related to the assignments in class. Or at least that is how it felt for me.

TA's: I did not feel like they were helpful or knowledgeable. Unfortunately, many of these TA's would not give concrete answers to binary questions. In many instances, they will answer with a question to your question that could have been answered with a simple yes or no. They were very impersonal. I hope it is not the same experience for future students.

Professor: He is barely involved in this class. I saw maybe a couple of comments from him.

Material/subjects: high level overview of all the relevant topics in analytics modeling. Linear regression, logistic regression, random forests, simulation, optimization - you name it, it's here, and you'll get a good conceptual overview. Really could not ask for more or better in this regard.

Exams: potentially tricky. Many questions are of the "which of the following would be appropriate" variety and you will be given, say, 4 choices and the answer may be 1, or 2, or 3. I found some of them to be debatable but there is no room for debate. I will say that on the whole I think the exams are effective in terms of gauging the student's grasp of the concepts, and I bombed one of them. The key is not to overthink it.

Homework: the syllabus says you don't need to know R and you'll be fine if you're willing to learn on the fly. This is inaccurate. The first few weeks are potentially very time consuming because depending on your background/experience you may have to do more than "learn on the fly" - you have to teach yourself the basics of a new language. If you majored in computer science and this is easy for you well congrats! Students with no R experience would be better served by being told "homework assignments will be in R, please be sure to have the latest R Studio installed and ready to go." Many posts here comment on the input/output aspect of the HW which is a small % of the grade. In light of the fact that there were something like 20, maybe 30, TAs I think it would benefit students if one of the three peer reviews was from a TA. I also think some sort of "evaluation of evaluations" would be helpful. It was clear that some reviewers were just being nice and giving 100's to everyone and it was also clear that some of the students are [EXPLETIVE DELETED] who wouldn't give more than a 90 to anything. I think if a student never gives less than 100 or never gives over 90 this should be factored into the grade (just examples). It's too easy to merely log in, check 90, and write "good work". That shouldn't be the case. I hope the staff considers evaluating the evaluations in the future.

Instructor: I'm in the minority here. Most people seem to really enjoy the guy's videos. I am not joking or being sarcastic - I have insomnia, I have trouble falling and staying asleep. I fell asleep during one of the PCA videos. Additionally, I found him to be very awkward on camera, he never seemed comfortable and never knew what to do with his hands. But really the most off-putting aspect.....and again, I'm not only in the minority but I seem to be the only one bothered by this, so maybe this can just be dismissed as an outlier.... the references to his consulting projects grew to be the bane of my existence. Does it establish credibility? Sure, I guess. But does it convey the message that he'd rather be somewhere else? Well, I see a lot of reviews here that say the professor never interacted directly with students. Correlation doesn't equal causation but might there be a connection???

Took this course along with 3 other courses in my first semester and this was my favorite! The course takes us through a wide range of ML models at a decent depth, different models every week, and make us think for real-life case studies near the end. The prof is very good (some of us call him a jedi) at explaining some technical depths in minutes.

There is a homework every week. Week 1-11 makes you code (mostly in R, some in Python, one potentially in Arena), which was first shocking to me as a new comer to coding, but the office hours laid out the basic parts of the answers. Homework are graded by peer reviews with emphasis on the extra depth of analysis that's presented (you can get a 90 if you follow the solutions in the office hours but you must put extra thoughts for a 100) which I loved! It lets you explore and expand your analysis. Week 12-14's homework don't make you code but ask you to think of a solution for a case study, which was a lot of fun!

There's a project that's similar to the case study homework but (ideally) longer and deeper, which was fun, too!

The midterm and final tests were MCQ, but the questions were non-trivial nor unnecessarily tricky. No silly mere plugging into formula or calculator. They were excellently designed to test your understanding of the concept. I just wish all MCQ exams are designed like this!

Overall a great course that has set a firm interest and fundamental basic knowledge in the field. I'm sure it'll be an asset for my future journey in data science!

Before I start I think I have to clarify between programs as a review. There were some people, I think from the OMSA side, that were very vocal against the fact that they had to program in R at the beginning of the semester with a few having borderline meltdowns on Piazza. I am from the OMSCS side so my review is from that PoV.

If you are from OMSCS the amount of programming is pretty trivial and yet the assignments are fantastic and mostly ML based along with some optimization/simulation modeling. I have been continually referring back to the homework's when developing some offshoot models at work so I think the content is awesome (random forest part was really interesting). I was very happy the course was in R as most OMSCS courses seem to be in Python and Java (great as well but its nice to learn other capabilities). I recommend taking it earlier in the program and pairing with another course as it will help build a nice base for ML, but shouldn't be over the top challenging.

There are homework's every week which usually just took me 1-3 hours. The office hours they give you 3/4 of the code so I think its worth waiting till Monday's office hours to start (due Wednesday). The assignments are peer graded, but don't worry it should play into your favor, especially if you talk like a programmer. There are the occasional mall cop peers in the group that absurdly give you a 75 on an A+ submission because you didn't do it like they did, but they take the median peer review grade so I never had an issue with this effecting my grade.

The only complaint I would have is the tests count for a whopping 75% of your grade. Additionally you can use written notes but this leads to a big luck factor - On one test I was lucky I made this one table that ended up happening to correspond closely to like 10 questions. On another test I had none of the equations written down and found myself guessing on 10 questions about which equation corresponded to which model. Your time is best spent watching the lectures weekly and then twice more before the test. Additionally, getting involved in following the Piazza discussion questions should help.

Overall I loved this course and its content - Definitely recommend taking.

This is a nice introductory course for ML. It touches upon several topics, ranging from Supervised Learning, Unsupervised Learning, Simulation, Time Series Analysis, Deterministic Optimization, and Design of Experiments. It also teaches you the fundamentals of data wrangling, which is extremely critical for any ML project.

The following three points outline why I liked this course -

1. Exposure to different ML concepts and R packages
2. A new way of thinking towards combining different tools to answer a business problem
3. The exams are tough, but they are designed to make the test taker think about the business scenario and how analytics can be used to solve the problem. The questions will make you think quite a bit

I have only one minor concern - Homework assignments add up to only 15%. However, one spends quite a bit of time, especially, if you are new to R

The TAs were awesome. They responded to every question within minutes of posting on Piazza. At the end of the course, several students thanks the TAs and Prof. Sokol for the experience.

Generally a good class. The homework is designed to force you to learn, but like others have said, be mindful of your time spent on it as the ROI is not very high. I felt the exams were very fair and not overly challenging, but they did make you think.

This class struggles in the peer review process though. It's such a toss-up on the quality of the reviewer that it's easy to lose value from any insight (or lack thereof) for each homework. I had reviewers give good marks when they were not deserved, and the inverse occurred as well, particularly during the writeup portions near the end of the semester. It is clear who does and does not know how to deduce from the grading recommendations, and it's also clear that some don't add any comments of value at all, which is disappointing. Shoutout to those that read each writeup and add thoughts other than "solid work."

Slack was far more useful than Piazza for this class, by the way. Piazza is just a UI mess...

It's just ok. Mediocre video lecture quality, nothing's very clear. You really need to find other materials/research other materials online to understand everything. This was my first OMSA course. Exam questions are sometimes very ambiguous, each exam has had questionable answers where you can dispute whether what's marked as wrong is actually wrong. Never seen the professor live, which was dissapointing. I did learn a ton, but it was mostly self study.

As an introductory course to the OMSA program, I think it's great. Prof is solid.

The key is accepting this course for what it is - an Introduction to Analytical Modeling. If that's what you're looking for, it's great. Each week introduces a new concept or two and provides a little detail on it. But there isn't much depth. If you're looking for more than that, you'll be disappointed.

Don't sweat the HWs. Simply completing the HW correctly is only worth 90/100. You have to go "above and beyond" to get 100, and how that is defined is at the whim of the students assigned to grade your assignment. I found the grading to be pretty inconsistent.

If you're struggling with the programming, just wait and watch the office hours, they provide a ton of help getting you started with the code. And while there are 14 HWs, only 13 are graded and you can drop your lowest two. So you only have to do 11. Also, the last two graded HWs are just essay style questions, no coding required. So if you're really struggling with the programming, plan your drops accordingly.

Focus on the lectures/concepts, take notes, and make a good cheat sheet. Some of the exam questions can be a little tricky, it's nice to have the security of the cheat sheet to help you out.

HWs are 15% of your grade and you can get 90s through watching the office hours and a couple hours of effort. There are three exams worth 75% of your total grade. It's better to spend an extra 4 hours studying per exam (for 12 extra hours total), than to spend an extra 4 per graded homework (44 extra). The extra HW time might get you 100s, but it might not. If you're pressed for time, focus on the exams. If you want to make sure you learn the programming material covered in the HWs, just be sure to review the solutions when they get released.

I found Piazza for this course to be pretty meh (at least compared to CSE 6040) and only checked it for the weekly announcements. Only posted something once and essentially had both students and TAs telling me I was wrong when their responses made clear they hadn't fully read my comment nor had they actually thought about it. But that's not a particularly surprising outcome on an internet forum...

Both OMSA and OMSCS students can take this course, so if you are interested in ML, I highly recommend that you take this course. This is an introductory statistics course, which covers varieties of statistics modelings such as SVM, k-nearest neighbour, k-means clustering, etc. These techniques often used in ML.

You use R and Python in this course and most of the homework assignments is done in R. I spent a lot of time on homework every week and it was pretty time-consuming. I didn't like that homeworks and course project re peer-reviewed and are graded by your peers. Also, I personally didn't like the tests as well. They were a lot different from homework problems and focused more on the concepts. You should review the lecture videos carefully. I used a reference sheet as a brief summary of each technique & formula and did well on the final.

• Testsx3 (Midterm1, Midterm2, Fianl): 75%
• Homework: 15%
• Course Project: 8%
• Syllabus Quiz: 2%

As many have said, the content is a high-level review of many analytical models. Very good introduction to the program. If you're coming in without a background in R then the first few homeworks will definitely require a lot more time than you might expect, but you'll learn a ton doing it.

The exams are fairly easy and don't require any math or programming, but rather an ability to select and interpret models/results. That said, with so many models, it's worth your time to really dive in and prepare, as many student seemed to feel the time allotted wasn't sufficient.

(OMSCS Fall 2020) I was totally new to ML and analytics and I am glad that I took this course. A broad survey course with wery well-produced, concise lectures. Dr. Sokol is awesome. If you don't have a strong foundation in probability& stats (Probability distributions, p-values), it will require a lot of groundwork to understand the lectures. Don't be fooled by the "easy" rating. It is very much intellectually intensive, if you are new to the concepts and want to do full justice to the learning process. But if you are just aiming for a B it can be done with relatively low effort, just be careful with the exams weighing 75%. Exams are well-designed and will test your understanding of the topics (There were some annoying trick questions, but much better compared to some other courses (IIS)

Time requirement is low except for studying for exams (3 exams ,weighing 25% each) and learning basics of R in the first couple of weeks.

I love this course. Content-wise, you don't learn as much, it's a fairly light course. Having said that, the amount of effort you need to put in is minimal, with generally very high ROI on learning per unit effort. Because of peer reviewing of assignments, there is generally little pressure to make everything perfect, meaning the focus is more on the learning. Once you're familiar with the coding the assignments are a breeze, so little busy work there. Exams are suitably hard, enough to distinguish those who put in the hard work but still spread out enough for someone who is already somewhat familiar with the content but didn't put in much effort on the course to do well enough. Lectures are short and sweet, information per unit time is high. Overall, an easy but enjoyable course.

A great survey course, where you definitely learn about a number of different areas that help you decide if/where you'd like to proceed further in the program.

Like others I came to the course with almost no R/programming background apart from about 15-20 hours of Swirl/youtube video cramming before the course, and like others the first 2 weeks were incredibly confusing re: the homeworks (if not the lectures, which were solid). Go to office hours and do more background research and just get through these first few weeks if you are new and feel overwhelmed; it gets easier. Also, understand that solid effort and clear explanation of what you are trying to do is usually more than enough to earn a 90% on the HWs.

I understand why others point to how 75% of the work is on the homeworks (worth only 16% of the grade) but I also understand how this informed conceptual understanding and why overall HW effort does not equal overall knowledge in the grading of this class and why the tests are disproportionately skewed toward that conceptual knowledge and not programming knowledge. As someone not proficient in R I appreciated this element of the course. The tests are tricky, but well designed. My only advice here is to lock down and submit answers as you go (after reviewing) rather than waiting to the end of the test, as many had proctortrack issues where crashes invalidated entire tests (in my case wiping the final when I had 20 minutes left and was just set to submit all my twice-reviewed answers). :)

I did learn quite a lot, and in the end I think this is a great introduction to the program.

Professor Sokol's videos were a pleasure to watch. You will get a very good intro into the world of analytics/data science which isn't just machine learning or predictive modelling per my previous misconception.

Some tips: Take the quizzes before your midterm and not immediately after you watch the videos, Use the transcript - will help you retain information easily rather than watching the videos multiple times. You can get the transcript by asking in slack group. Do a basic R Course on Udemy or some other platform. Knowing python OOP will help; learn packages Pulp and SimPy if you can. There are two assignments for which you will need Pulp and SimPy. Do not follow the advice given in one of the reviews below that says avoid trying for 100 on homework assignments. Difference between 90 and 100 is just 0.2 in the weighted score on each homework but these fractions can be very important as it is extremely difficult to score points in two mid terms and final as the some questions can be ambiguous and some very tough. My weighted score in this course is ~87.8 and I was lucky for the curve to be in my favor and I got a A. So, try for 100 on every assignment and project!

Overall a good, but not amazing, class.

This was the second class I took in OMSA after CSE-6040. Decent Python background, not as strong math or R background but didn't come into the class with zero knowledge. I ended the class with a solid A.

I definitely learned a lot, the quality of the lecture videos is generally very high. Concepts are explained intuitively and delivered in a manner well suited for online learning i.e. short, dense videos with clear objectives and explanations.

The exams are very well structured, they reference material almost exclusively from the lecture videos so I found studying for these relatively straightforward, though that wasn't the case for everyone and for some the exam experience was a little jarring.

For those above reasons, I'd rate the class as good and recommend it as an early class if you are an OMSA student or an elective if OMSCS.

The class does have some flaws that could be improved upon, though.

1. As others have mentioned, the effort to reward ratio on the homework is not optimal. To be fair, Prof Sokol describes the homework as a learning tool and the low weight assigned to the homework grades is likely due to this. But some homework assignments took an inordinate amount of time, and sometimes it felt necessary to add in some extra busy work just so your peers would grade you the 100 for 'above and beyond' work.

2. The pacing could be improved upon. This may be more specific to the summer schedule of the class in which some weeks crammed in a very large amount of topics but nonetheless. The final three weeks of the course are dedicated to vague case studies where you write and describe models you use in the situation given. Since I had not used any of the two free homework drops provided I skipped the last two of these. It felt like this time in the course could have been used to spread out some of the denser weeks, or dedicated to a more in-depth overview of R at the start of the course.

3. TA quality and general Piazza conduct was poor overall. The head TA posts weekly updates that are informative and helpful to the class. But I found in a majority of cases the TAs to be reluctant to engage in student discussion, more focused on 'resolving' a piazza thread than actually helping students learn. After my first post where a TA misread my question and answered incorrectly at first (but fast!), and then ambiguously when I restated the problem once more I gave up on Piazza as a source of meaningful discussion, which is a shame. The other two classes I've taken had very strong Piazza forums, fortunately for this class the Slack community was very strong and provided a much better source of student discussion.

I did not attend Office Hours once I realized they were essentially providing pre-written R code to help with the homework and little more.

Final summary: Good class, learned a lot, would recommend but it could also be better.

The course suffers from lack of oversight. There are literally hundreds to 1,000 students taking the course (if TA LinkedIn profiles are to be trusted). The professor held exactly 0 of the office hours. Office hours consist of various TAs going over code which can be used in homework and not discussing the material covered in the module. At least not very well. On more than one occasion the TAs gave incorrect information when students asked questions (I would say less than 10 times during office hours). One particularly bad session consisted of a TA looking into the camera and reading the code. Not screen sharing, not split screen, just reading the code.

I found the course fairly easy, it was simply tedious in how much information you need to write into homework to ensure your peers think you provided "a deeper solution than expected" the requirements. Answering all questions correctly will result in a 90, to earn a 100% you have to do more than what is required. This should either be described differently or changed. If it is required to describe the theories underlying a task in great detail just make that part of the assignment. It is akin to saying to earn the full 40 hours paycheck the person really needs to come in and work 50 hours to show how dedicated they are.

You will not know what your overall grade will be until after the class is over. You will not be told the cutoffs for particular grades. I've had classes where grade scales were changed over the duration of a course, sometimes you do that to ensure a course is fair. But never just leaving students in the dark. It is bad policy.

This is a mandatory course, it covers a large amount of information, most students were well prepared for it. The workload was not too much, the tests seemed fair, more focused on showing an understanding of theory than particular coding application (which matched how the course was described). But there are annoying issues for a course, which 1,000+ students, $500-700 per student, so it is bringing in more than a half million per semester, maybe as much as $700,000, this should be a much better course.

I did really like the course, but I don't think it's an 'easy' class, especially if you do not have stat/prob and programming background. I did get an A in the class (before any curve) and have gotten 100% on all HWs and Project but this required a lot of work for me.

The course covers a lot of materials (at surface level) and expects you to do some more learning outside of the class on your own. Sokol somehow makes what could be tedious lectures actually interesting and entertaining. TA office hours are really helpful for the homework and this might actually save you a lot of time while working your HWs (the workload is so high for me b/c I did not discover it until the middle of the semester). As this is pretty effing dense and moves quickly, I would not recommend taking this course in the summer.

Tests are... well, they are still a mystery to me. You just need to get used to understand what they are trying to test you on and their test style to do well. Also, in my opinion, as there are many versions of the test questions, you do need luck on your side. Even when you're being tested on the same / similar concepts, I think someone more trickier than the others.

For the 1st test, I studied the most, went in pretty darn confidently (after studying super thoroughly) but did poorly. I think this really had to do with me not interpreting their test questions correctly or understanding their test style rather than really showing how much of the material I understood. My soul was crushed for weeks... 2nd one and the final, I did really well (but again, I feel like luck really played into effect) despite not studying for it half as much.

I have a lot to complain about their grading scale. I really did not like how the weekly HWs take so much time and effort to do but are only worth 16% of your grade. I feel like this should be bumped up to 30% or so. Also, I do not like how for the HW and Projects, if you get everything right and do it correctly, base grade is 90. You have to go above and beyond to get that 100%. No, you don't get an extra credit for doing that. Honestly, it's easy to get a 75 or 90, but going from 90 to a 100 is a lot more work for 0.2% of the grade lol. Also, there are some assholes peer grading who will leave a feedback saying how great the work was and give you a 90 sometimes. Project was pretty neat, applying what you've learned in more practical situations.

Anyway this course was rough but you'll really like it and will learn a lot from it I think.

This was my first OMSA course and I thoroughly enjoyed it. Professor delivered the content very well and it builds slowly from python boot camp to more complex python skills. Knowledge of math, data structures and some programming basic will help. I believe spring 2019 was when there was an incident during one of the mids when some students chested. Our final formay was changed from mid terms. Mids had around 2 days to complete whereas final had 3 hrs. Neither were proctored and all exams were open book. I heard students had a tough time with exams with limited time during later semesters. I can totally see it. I did well because we had so much time to complete and I am not the sort who can come up with code, debugg it and make it all work quickly.

There are Weekly assignments, all based on completing python notebooks online, mid terms and final, all coding based in python. I hadn't used python before and at the end of the course was able to write decent code (even for my work projects). I found this course extremely helpful. Scored 92%.

This is a very good course. A perfect starter course for OMSA program. It provides a well structured and well delivered overview of ML models. You will get to know a lot of things in this course but at a high level. It doesn't go into details of math and proofs a lot. But knowledge of math and stats/probability concepts will be needed to do well in exams.

Weekly coding assignments in R keep the course busy but they weigh very less in the total course percentage. Project is descriptive, no coding needed. Mid and final are well drafted and would have questions with all levels of difficulty. I had not scored well in the mid term because I didn't know how to study. Key is to re-watch videos at least 2-3 times, reread all slides and take really good notes. Some questions on the exam relate to more higher level math and details than covered in the course but I guess those are the differentiator questions. I manage an 86% finally with about 10 hrs per week on avg.

Good course! Took it as the first course in my OMSA journey. Gives a good overview of the various modeling paradigms. I liked the weekly home works as they kept us on top of things and in schedule. The office hours were concise and gave a starter code for the homework. I liked how the exams stressed on concepts and made sure that through three exams the concepts kind of become ingrained.

What could be improved is the peer grading randomness. Some times the best of our homework attempts might fetch a much lesser score than the homework which probably wasn't the best.

Also if you don't have some programming experience, please do consider spending a little more time than 10 hours/week.

All in all, a great course run by Dr. Joel Sokol.

This is a good beginning course for people.

Pros: Theoretical, Focuses on concepts and good for an understanding of concepts.

Cons: Weekly HWs. I prefer more dense Hws every 2-3 weeks. I forgot to submit quite a few of these HWs and it ended up costing me. Not much programming. I took this course because i wanted to learn R. But hardly any programming at all.

I thought that the professor and TAs did a great job, especially in light of the coronavirus challenges. There was a homework due each week that is peer-graded, and you have to do 3 peer-grades each week. Once you get used to that, it isn't too bad. The majority of the class is focused on modeling techniques, and the last 3-4 weeks are focused on case studies, as is the course project. I found the questions on the first and second midterms to be tricky, but the final seemed easier to me. If you did your undergrad at Tech, then you'll find this class to be in line with the setup of many others. For those who didn't, don't be surprised to find that assignments are due the same weeks as the midterms. This course mostly uses R, which took some learning, as I use Python regularly, but it wasn't bad. All in all, I enjoyed the class, and found the material to be applicable and well developed.

I took this course as the first course in my OMSA program , felt this will be the right stepping stone in to the masters. But I felt the course was trying to go over a vast majority of the topics in a short span. The biggest issue was the peer graded assignments being a hit or a miss based on the eyes of the reviewer. The subject also did not have any text books or reading materials . So had to always go back to the videos and lectures to take notes which made it a painstaking process for exam preparation. I hardly got proper responses from TAs for the questions raised.

This was a difficult first CS course for me as I am completely new to almost every concept covered (R, the models, Python, etc.). While I really enjoyed it, it was a stretch for me.

Pros

• I consistently earned 90%'s even though I struggled to get through most of the assignments. If you do a decent job and put out effort, you'll get 90s and the occasional 75%. I thought it was more than generous.
• I was able to learn R while doing the assignments. It was like being thrown into the deep end of the pool, but YouTube became a close friend and I fumbled my way through.
• Your two lowest HW scores are dropped
• The project (9% of your grade) was fun and very easy. If you're creative and have a passing understanding of the concepts and business needs, it'll be easy.
• The lectures are short, but interesting and hold one's attention
• The tests and the course are graded on a curve (by some miracle I earned a B, when I clearly did C work IMO)
• I really put in minimal time (on average about 8-12 hrs/wk)

Cons

• Mid-term tests are ridiculous. They are the most abstract elements from each lecture. I did okay on the first one and got absolutely rocked on the second one.
• The final, in comparison, was very easy and common sense driven. Again, if you have a passing understanding of the material you'll be fine (and will probably have over studied after the 2nd mid-term)
• The lectures had almost nothing to do with the assignments. The lecture would cover the concept of "lettuce", then the HW assignment would be "make a Thanksgiving dinner". Nothing seemed to connect.
• Some of the TAs were obnoxious and pompous. As soon as I saw some (one in particular) lecturing, I would just signed off because I knew it wasn't going to be worth the time and the browbeating (cheers to some of my classmates who were relentless in their questioning! :D). While this particular TA knew his stuff, he clearly has no skill in teaching (I say this as a professor myself).
• I don't think Dr. Sokol interacted with the class...ever (I mean, he was literally just the name attached to the course and the talking head in the videos). I'm sure he's a super busy guy, but it would have been great to have had some passing interaction.

Overall, I'm glad this was my first course as it gave me a good baseline of what to expect. I rated this course as "Hard" based simply on the exams. Weighing all factors together, it was probably closer to a "Medium" (but I want to forewarn others to prepare diligently for those exams!).

I took this class as an OMSCS elective. Overall a well-paced course; not too hard yet not too easy. This survey course had some small overlap with ML4T CS 7646 which was nice, and i'm sure it would be a decent precursor to ML CS 7641 (which I have not taken).

1. There's a weekly HW that is peer-reviewed... so just aim for a 90% which is feasible. Getting a 100% requires your peers to view your work as "above and beyond" which is subjective, so don't kill yourself over it unless you really enjoy the content. There are some HWs that have optional sections, if you do those you have a good shot at the 100%.

2. The final project was quite open-ended; I would gauge it at about the effort of 3 or 4 HWs combined.

3. The exams were a bit stressful, but as long as you re-watch all the videos and make your cheat sheet you should be fine.

I came to this class from OMSCS because I couldn't get a seat in CS-7641 (ML). This course is not an elective in the CS ML curriculum, though it probably should be. It's a gentle introduction to general ML topics and much of the material intersects with CS-7641/7646. At first I wasn't happy about having to learn R, and the effort did take up to 20 hrs/week through the first midterm. It eases after that, with the final weeks being close to no effort. As for R, I really don't like it, but it is a useful skill. It allows quick prototyping and handles data sets much better than Python. Python is the better choice for really large data sets and deployed solutions.

The Bad

• There are 15 homeworks/assignments, some of which take a great deal of effort. In the end, they only account for 16% of the grade. You will spend 80% of your time in this course earning only 16% of the grade.

• There is no assigned textbook, though the course covers material that's been around for decades. This makes it impossible to prepare for the 3 tests by doing any more than re-watching the videos.

• Two tests and one final, accounting for 75% of the grade. They released sample problems for the first test but nothing else for the other two. There was no way to adequately study. The tests were hard too, with many tricky and gotcha questions.

• The homeworks are peer reviewed. Feedback is mandatory, but most reviewers will only leave a comment saying "good job" and assign an arbitrary grade. Full effort only gets 90%, and there is no consistent guideline on how to earn the other 10%.

• This is the first course for the analytics students and many of them were completely culture shocked. Too many dumb/redundant/whiny questions on Piazza. My favorite was the people who were triggered because they were not allowed a bathroom break during the proctored exams.

• It's not possible to work ahead because homeworks and lectures are unlocked per the weekly schedule.

• There is a curve at the end, but Dr. Sokol keeps the distribution to himself. You just get to see your final grade.

The Good

• edX, much better than Udacity, uses NLP to produce linkable video transcripts.

• Superb organization, the course runs like an optimal algorithm.

• TAs Angela and Chase really knew their stuff and kept the course running along nicely.

• Similar in quality to Andrew Ng's ML class.

Overall

Prepare to spend 20+ hrs/week until you master R. Most of your time will be spent on the homeworks, which will make little difference in your final grade. This is an easy class otherwise, but the tests are hard and you will have to earn your A. Use R Markdown to prepare your homeworks. Dr. Sokol seemed to be active behind the scenes but his presence wasn't felt much.

Pro:

1. this is an easy class that more like undergraduate level, which provides a basic and wide range of topics

2. fourteen HWs (probably designed by the professor) are actually pretty good, which helps refresh the knowledge from course and R.

Con:

1. very little math, the professor tries to teach students "common sense" of the analytical driven model.

2. three exams are so badly designed (probably by TAs). Very nerdy and meaningless question without business common sense (like one multiple-choice its corrected answer is to convince CEO to take an obviously bad decision because of its "mathematically beautiful". Advice: just don't think this is graduate-level class and don't use any business common sense.

3. zero interaction from Professor. TA is not knowledgeable.

Overall, okay class, u would learn a few from videos and HWs.

This is my first class in the program. I do not have any statistic or computer science background. My only related knowledge is college math almost 10 years ago. I found this class not difficult. Most of my time spent on searching R documentation to figure out how to use certain function in R. The course is well organized and I do learned some. One thing I learned the most is google how to solve certain problems by myself. However, there are lack of mathematical explanations in some of the concepts or models. I do get a sense of what data analytics do, but I am not confident in handling a project. I am kind of disappointed that the final project is almost a writing project, which did not include any coding or actual problem-solving.

Coming from OMSCS with weak background in statistics and very little knowledge of R, I had some challenges with this course, but in a good way. I liked this course very much and found it well organized.

Advice to both do well and learn a lot:

1. Strive for 100 on every homework assignment and on the project. You can pretty easily get 90 on the homework, but 'above and beyond' is where you learn. Also, the homework and project are worth 25% - the same as one exam - and I found (the hard way unfortunately) that 1 point extra on the homework is way easier to get than one point higher on an exam.

2. Use R markdown to neatly combine code, writing and graphics to produce a paper you're proud of and also impress your peer graders. Take the time to learn R as you go if you don't know it yet, and learn to make neat plots and stuff.

3. Study really well for the tests. They aren't out to trick you but they are long and thorough and you better know the material of the lectures cold. I needed to supplement the lectures with further explanations on stats - like hypothesis testing, p-values, confidence intervals, probability distributions etc - and I watched other lectures on models to get a better understanding.

4. Leave plenty of time for writing up a good cheat-sheet. I estimate it took me 8 hours each time to review material and write out the sheet(s).

5. The UI of the exams is not so friendly. So make sure you submit everything at the end by searching for "0 of 1 attempt" (and don't find it!), make sure you click submit for every question.

6. The head TA and other TAs were great. The office hours are very helpful for getting started on the homework. Homework solutions are posted too, so that's another great way to learn. The TAs promptly respond on Piazza and Slack to all and every question. There was only one nasty TA who was sarcastic and unhelpful during his office hours.

7. The material of the course is fascinating and immediately applicable. By the end of the semester, we were all modeling COVID-19 of course.

I'm in the OMSCS program and I took this course alongside ML, which I definitely preferred. I was interested in taking a course in another discipline to see how other programs were. Overall, the course wasn't difficult and was a light load which was nice when combined with my 20+ hour workload for ML. But I would say this was my least favorite course in the program so far and honestly regret choosing to take it instead of another OMSCS course but I think that's mostly due to the fact that I didn't learn a ton of new material since most of it fell under ML/things I had already known. I did get an uncurved A in case that matters.

Here are my general complaints:

• This has been mentioned in reviews below but the homework is only weighted 16% while each of the three exams is weighed 25%. This disparity is strange but I think it's due to the fact that there are no "right answers" for most of the homeworks.
• Each exam frustrated me specifically the fact that it's mostly 'select all that apply' questions. I think the contents of the exams were a good test of the application of the material rather than testing based on just concepts. I've never had to use edx to take exams but the UI is terrible. You have to individually submit each question (i.e. there's no overall submit button and it doesn't auto-submit by the end of your time) and once you submit, you can't change your answer.
• Piazza - it's a mess. There were many duplicate posts and it was evident that posters didn't bother to do a quick search before posting...There were also so many questions that could easily be answered by just reading the syllabus. I'm not sure if it's because this is an intro course or what but as a fifth semester student, this was insane. I think the quality of the course would be greatly improved if the TAs discouraged this behavior. Following that, I found my peers to be frustrating at times. The general maturity level of students in this course was not up to par with what I've experienced in my other courses.
• Peer reviews - generally, peer reviews were alright. But a lot of reviewers put in no effort or were outright aggressive (and incorrectly so...). I'm not a fan of peers grading assignments for grades. I get the value in seeing different approaches by reviewing your peers' assignments but I think Dr. Joyner's courses do this a lot better. In those, rather than having peers grade assignments, you have participation points for reviewing peer assignments to actually get some feedback back instead of having people just give you a score. I'm not saying this as someone who had a low homework score (I got a 99%), but as someone who gives thorough feedback who is frustrated by the lack of effort from many of my classmates. I can't blame my peers for this as I believe this is just how peer reviews are structured in this course (i.e. you can give a grade with no comment).

While I didn't really like this course, there were a few positive aspects to it:

• It didn't take that much time. You can realistically finish each weekly assignment within a few hours. The first one took a slight bit longer because I'm a python/C programmer and hadn't touched R before this course.
• There were two homework drops which was nice (while I didn't use them, I did get to drop two of my 90 scores woot)
• The Head TA, Angela, was great at posting weekly updates that reminded you of what you had to do each week. Most of the TAs that hosted office hours were great (specific s/o to Ozan, Kristina, and Pin who were phenomenal) although there was one that was unnecessarily snarky.

If you choose to take this course, here's some advice:

• Assume that the exams aren't trying to trick you. I did the worst on exam 1 because I thought the exam was trying to trick me after reading reviews here and it wasn't. I took a more trusting approach to the other exams and scored much higher on them.
• Avoid piazza if you can
• For me, a lot of the value in cheat sheets was making them. I rarely used them during the exam since I had retained information by writing them on my cheat sheets.

If you need a light course to balance a tougher one, you can take this. Just beware that there's going to be frustrating aspects to it.

6501 is generally regarded as one of the more well designed courses in OMSA. I can see why, though if this is one of the best courses it makes me a bit concerned for the rest of the program. Maybe I’m spoiled because I’ve been in good grad programs in the past, but I wasn’t as impressed with this course as some others.

The good

• Lectures are well produced, so I was never struggling to hear or see anything.
• Syllabus is thoughtfully laid out. For the most part topics follow naturally and the order makes sense.
• Homework assignments are generally reasonable and let you get some good practice going through all the steps of each model the course covers.

The bad

• Peer grading can be hit or miss. TA’s typically remedy poor peer grading within a few days when asked.
• Piazza is a disaster. There is no editing of questions, no defense against duplicates or garbage posts, no expectations for what people include in posts asking for help. In a class with 20 people this might not be too bad, but with 5000 it makes Piazza borderline unusable.
• Homework does not prepare you for the exams. I don’t think the exams are necessarily unfair or that they exceed the scope of the course material. There is just a tremendous mismatch between the exams and all other assignments.

The Ugly

• No professor interaction at all, TAs are reluctant to engage in a discussion or provide resources. I can’t tell you how many times I read thoughtful, relevant questions on Piazza only to see the TA response merely tell the student to re-watch the lecture videos. I might accept this from a coding boot camp, but not a renowned university.
• The final is 3 hours with no breaks. Frankly, it’s ridiculous to not allow a bathroom break. Even the GRE gives breaks.

If you want to get into data science this is the first course you should take in the program. Survey of many modeling/ machine learning techniques. Never goes very deep into any one topic. Homeworks are pretty straightforward. Having to deal with them being peer-reviewed is a pain but not the end of the world.

Where you have to watch out is the exams. They are tricky and expect a level of familiarity much greater than you would expect from lectures. All the content from the exams is based on lectures, but the level of depth is greater in the exams. They give you a good amount of time for the exams so use it!

I liked this class. It was the expected level of difficulty for a graduate class. It gave me the mathematical explanation I was looking for in relation to a broad set of algorithms. I was new to R but have a Comp Sci background. I thought the exams were fair. I did spend significant time on the homeworks and studying for the exams. I got an A.

With the rise of sites like Coursera, EdX, & Udacity, I have taken dozens of courses online. This access to courses from Universities at every level of prestige has significantly raised the bar for what is an acceptably prepared, taught and managed course. I strongly believe that this will begin to weed out professors who spend most of their time on research, who are not constantly updating and improving their course content (e.g. lectures, homework, etc.) and are not deeply engaged with the 'teaching' experience. That said, with respect to IYSE-6501, I have definitely taken worse courses, but compared to the best classes out there, at this time, available at a fraction of the price, this class felt like an introductory undergraduate course, a General Education requirement, at best. Although the content itself is arguably very advanced, the level at which it was taught, the detail provided, and the challenge presented by the homework and 'course project' were insufficient. I spent more time watching videos and reading material outside of this course to make sense of what was presented in videos that rarely even add up to 30 minutes of lecture for a given week. The average undergraduate course has at least 2.5 - 3 hours of lecture per week, for reference.

Peer reviewed homework is a controversial subject, but should not be the only or primary way that homework assignments are graded, in my opinion, particularly in a course where concrete questions are not hard to produce, and when there is a cornucopia of software with examples out there to better engage students in programming assignments. The amount of detail and challenges presented by the homework left too much room to the students to 'fudge' their way through. You can argue that it's up to the student to get what they want out of the course, that they will get what they put into it, but when the student has to put in 90% of the work compared to the 5% provided by the course itself, and the other 5% by their peers, I'd say this is a gross imbalance, and rather lazy way to teach at this level. I don't mind working hard, but I'm frustrated by the number of times I find myself doing well in courses 'despite' the instruction or course structure, instead of with the aid of the instruction and complimentary materials. Often, this is what a textbook will provide, an angel when lectures are otherwise not particularly helpful. Textbooks also provide more homework problems, and a much greater level of detail. I don't mind paying for textbooks, if it means that I'll actually get more 'content' for the extra 50 - 200 dollars.

I have to compliment Joel for his presentation as a person, and I honestly like him as the voice behind the content, and I get the feeling he is very passionate and committed to the program and his students. But with regard to the material in this class, the slides were too 'static', the videos were much too short, and the content too brief, even for a whirlwind tour of dozens of very complex subjects. I think I'd recommend for starters that the faculty look at courses like the Machine Learning specialization from the University of Washington on Coursera, for what I feel should be the average level of instruction to expect from a graduate course, and these classes are even more affordable than Georgia Tech's.

I'm sure I could say more, and I likely will, at some point. I sincerely hope that Georgia Tech doubles the staff employed to create content for their courses. I applaud them for leading the charge in providing an incredibly affordable, highly 'reputable' graduate program, but I'd push that they lead again, and provide both the most affordable program, in addition to the best organized and taught courses in the industry.

I have mixed feelings about this course. I like the topics in this course, you can learn about many interesting modeling techniques and you can learn a lot by doing the homeworks. Notice the word "can", because I think this is a course for self learners. I think most lectures are very superficial and most of the time you will need some extra material to understand the topic of the lecture... I mean, some lectures are ok, but most of them are superficial. The TAs were not as helpful as I expected, but there were some exceptions. Overall, I learned and I liked what I learned, but I think the course material could dive deeper into the subjects.

This is an excellent introductory course for analytics concepts and popular models, homework is fun and a little challenging. The lectures are more like general guidance instead of step-by-step instructions. Exam questions are well-designed, students are expected to learn things fast and have a true understanding of the models instead of just memorizing concepts.

As many students have mentioned, the course is a survey of analytical models, and running ML models in R like classification, clustering, and regression. The homeworks are the heaviest part of this course, but thankfully the TAs were amazing. I highly recommend at least watching the office hours that are conducted by the TAs if you have never programmed in R.

Prof Joel Sokol is amazing with his lectures which takes a high level view in a way that piques your interest in the other courses in the OMSA syllabus such as regression, optimization, and time series models.

The exams take up the highest percentage of the grade which can be tricky but definitely manageable. You are allowed 1 sheet of cheat sheet (handwritten only) for the midterms and 2 sheets for the finals though I haven't really used them since the exams are less about regurgitating equations and more on applying analytical models to specific scenarios.

As it has been mentioned by many that the amount of time taken to complete the homeworks are disproportionate to the time taken to prepare the exams, I respectfully disagree that it is a bad idea since Prof Sokol has mentioned at the beginning that the homeworks are focused on learning and this is not particularly an R course. That being said, even simply by following the example codes given in the office hours and doing some minimal tweaking would easily earn you the 90% grade for the homeworks. However, I recommend going deeper even if it is a diminishing return of 10% more in your homework score relative to the time taken since it will definitely help you in your analytics career in the long run and definitely solidify your proficiency in programming to a certain extent.

For those who are really afraid of programming and feel you need a little more hand holding with regards to learning R, I would highly recommend taking "The Analytics Edge" by MIT and also on edX. It teaches many of the concepts in ISYE 6501 but with a greater focus on guiding you step by step with R syntax and packages.

Definitely take this as your first OMSA or MM course, you won't regret it!

I took this course in my 3rd semester. I take only 1 course per semester. I already took CSE 6040 and MGT 8803. The work load was comparable to CS6040 . The course uses R as the primary language. I never had used R, so before starting the semester I looked at the DataCamp R course, just to familiarize myself with it But did not spend too much time on it, maybe a few hours. The homeworks take a lot of time, if you want to take the whole points. There are 14 homeworks , 1 per week, that you have to submit every week. It then gets assigned to 3 peers (you have to grade 3 peers as well, by the next sunday). The grades can be 0, 50,75,90 or 100. If you have everything right you get a 90, but if you do extra work you will get a 100. 2 lowest grades can be dropped at the end. So only 12 homeworks will be counted. I spent a lot of time on the homeworks during the semester , specially up to week 7-8. Then it gets easier. Homeworks have 16% of the final grade though. I had 2 homeworks at 90 and the rest at 100 so I ended up losing only 0.5 of the final from the HWs. There is a final project , no coding, but finding a topic from the ones that are presented to you and writing a few paragraph on how you can use different models you learned in the semester to solve the problem. There are 2 midterms and 1 final , each having 25% of the final grade. The exams are proctortracked , 90-90 and 180 min each. You won't have any problem with time (if edx and proctor track work fine), a lot of people this semester had problem with that. There is no coding involved in the exam , and it is only based on the concepts. So just watch the course videos and try to understand the whole concept fine. In the first midterm my grade was 79 , but for the second exam I watched the videos twice and I read the questions really well and got 100. The final was cumulative of everything and I got 92. The TAs of the course are amazing. Very responsive , so you are never waiting on a response from them. So helpful! The office hours are really helpful for doing the HWs, as they provide the base of the codes and you can ask your questions, so even if you can not attend them, try to watch the recorded version.

All in all , I think this was a very good course and maybe it would have been better for me to take it in the first semester instead of CS6040.

What I would recommend to be changed is the proportion of the HW grade toward final. It pretty much took most of the time in the semester but it was only 16% of the final grade.

This was a really good course to be taken as a first. It touches the basics of all types of models and gives you a good understanding. Dr. Sokol was very fun to watch but homeworks can be very daunting. Office hours helps a lot. Exams were tricky. You need to go through the videos multiple times to catch all the information and make sure you're not missing anything as every tiny bit of info is important for the exam.

This was my first course and I really liked it. it gives you a good perspective of different models to be practice in other courses with good enough depth

Great class for everybody who wants to understand what Analytics can and cannot do! It also provides opportunity to learn how to do some of these things by hand, which is very nice.

I totally enjoyed it, and it was very easy. It's a bit sad that it's over.

This course started strong and required 3 hours of work a week, then quickly devolved into a joke the last 3 weeks where literally I learned nothing and spent 30 minutes on my homework assignments. If you are already an analytics professional, don't expect to learn much new the last few weeks from the cases. If I could do it again, I would probably tell Georgia Tech to let me opt out of this course and take something else instead.

I feel like you should only take this course if one of these two criteria fit you: a) You want a super easy credit to balance out a harder course b) You've never been exposed to any analytical modeling period

I had assumed we would go at least somewhat deeper into some of the techniques, insteads its just a whirlwind tour from concept to concept with very minimal guidance. This is literally a survey course that I would expect at a community college or an undergraduate level institution, not a school of Georgia Tech's rigorous caliber. That said, maybe this is intentional to ease novices into the analytics field. If so, Dr. Sokel did an excellent job, especially focusing on high level concepts on the test which were very fair.

The one unit I did learn something was the optimization one, where I spent an inordinate amount of time trying to debug my PULP code but then gave up. Still, I now mainly use Python for my coding at work, and refreshing my R was pretty nice.

Its interesting how much I loved the python course even though I spent 10 hours a week on the homework problems + tests , but felt extremely ambivalent about this course even though I spent 1 hour a week on average.

I liked this course. I thought it was a good introduction to many of the analytical techniques that are covered in other OMSA courses. However, the course packs a lot in especially given the shorter summer semester, so be prepared if you take it over the summer. The homework was open-ended in that you could spend a lot of time on them and learn a lot in the process, or you could do the bare minimum. Either way the HWs are only worth 16% of your grade in total, but they teach you a lot in preparation for the exams. The exams were very tricky and really tested your analytical ability and ability to reason through the problems in context of the course material covered.

Overall it was a good first course to be introduced to many different types of models and such. Unfortunately, I was very interested in the homeworks and put way too much time into those rather than studying for the exams; homeworks are 16%, exams are 75%. I have no complaints about the course (office hours were FANTASTIC, I listened in afterwards and they pretty much showed you how to set up the homework for a 90 every time). However, the exam format was something I have not seen before (I come from an engineering background where either I am right or wrong, nothing wishy washy about it). Honestly, each question could have multiple answers but you have to read VERY CAREFULLY, otherwise you will most likely interpret it wrong like I did on the first midterm. That's just my opinion since I am not used to those types of exams.
I have coded before, but never R and the TAs were very helpful in giving starter code and learning week to week. Dr. Sokol is clearly very passionate, it comes across in the lectures. However I usually had to watch them a few times; one time to take screenshots of the notes and another to actually listen to the content.

My advice: watch the TA office hour videos before starting the homework (they give the starter code), take ALL given time for the exams to read each question very very carefully. Prioritize studying over homework if you need to pick one. Don't forget to peer grade :)

Some notes on time management and grading- I took this class in the Summer, and the compressed schedule means there's quite a bit of work each week, and it's pretty front loaded. Homework can be time consuming, but not a high portion of your grade. It does not take much effort to get a 75-90% on homeworks, but will take a lot of effort to get a 100%. The difference in your final grade is negligible, and the last 3 homeworks involve no coding, where it's much easier to get a 100% with far less effort. You get two homework drops, I advise not doing the homework the week a Midterm is due and instead cramming for the midterms, which will amount for 50% of your grades. If you are struggling with time, prioritize studying for the tests over doing the homework!

The midterms are heavily conceptual, and your performance on the homework has no bearing on your ability to regurgitate the lecture material for the closed book exams. The exam questions can have some tricky wording, take your time and slowly work through the exam.

On the actual class- Now, some comments on the actual content of the class. I thought the material was a great survey course, touching a lot of different analytic techniques and how to apply them to various problems. I would often watch additional videos on youtube or do some reading on the material as the lecture material was pretty short on these topics. I did enjoy doing the homework and learning R, but was frustrated how little the homework/projects counted for my grade, and how it didn't seem to have any bearing on my performance on the midterms (I figured this out in the back half of the class).

I generally agree with the other reviews - good class, probably a good first class since it gives you an overview of electives you might want to take/what the program is all about, etc. If you don't enjoy this class, I'd seriously rethink being in the program since a huge chunk of the electives you can take are effectively a more in-depth runthrough of something you already learned in this program. If you don't like intro to analytics, you probably won't like an entire program devoted to it.

The lectures are interesting and well put-together, you can tell Dr. Sokol is passionate both about this topic and helping us learn. There's a lot of material to watch but it's generally entertaining. His occasional corny joke helps!

Most of the homework is done in R, with a handful that are writing short essays ("how would you attack this problem") and two that are in Python/Arena. Unfortunately, the lectures don't teach you how to use any of these tools, whatsoever. So you have to self-teach (though the office hours are helpful), which is a good way to learn but not why you sign up to get a degree, you expect some teaching more than "figure it out for yourself". If you don't know any R going in I'd advise doing some self-study before the class begins.

There are three exams, which have very little, if any, overlap with the homework but feed heavily from the lectures. The questions are very subjective, like "which model is best to use in this situation" and there isn't an obvious answer. Many of the questions are multiple choice and partial credit, like "which of these models could or couldn't be used for this purpose, select all" but not all of them follow this pattern. Personally I felt the exams were fairly tricky for this reason, it's not like a math or coding problem where you're clearly right, or not. Additionally we weren't given any practice exams to help prepare, I really wasn't expecting the format of the first exam and then I gradually did better the next two.

All in all an enjoyable class, as long as you do the work and understand the material it shouldn't be too hard to get at least a "B" in but quite a few people struggled too much on the exams to wind up with an "A", including yours truly.

This course gives you a quick introduction to a whole range of analytics topics. Nothing is done in-depth rigorously but that is not the point of this course. If you enjoy a topic and want to learn more then you take the elective later.

The homework is worth very little of your grade but is engaging and enjoyable. The final 3 were completely different with no coding and personally I didn't find them very engaging. The first 7 are based in R and whilst you can learn R on the fly I would highly recommend studying a few chapters of "An Introduction to Statistical Learning " and doing the exercises which are in R. I did this and found it incredibly beneficial.

Most of the code is given to you anyway in the office hours but if you write it yourself you will learn so much more. Homework is graded by 3 peers and if you answer what they ask correctly you get a 90, if you go beyond the solution you get a 100. Some weeks I spent hours and hours and got a 90 and then some I completed quickly and got 100. The homework counts for little anyway.

Exams were multiple choice questions which tested all the material in the course. I thought they were very well written and fair. Other students complained but I actually enjoyed them. They take no time compared to CSE6040!!! :)

One thing I will say is that taking this in the summer was a challenge as the content was squashed into a smaller amount of time and none was dropped. There was one particular week where I got overwhelmed and it wasn't fun. Otherwise I managed ok. Most of the TAs were good enough in the office hours and fairly responsive in piazza, although not great. Piazza and slack in general weren't as useful as CSE 6040 although this could have been because of summer.

Joel is an excellent professor. He is passionate about what he teaches and the assignments are all based on R. The course covers a lot of materials ranging from time series forecasting to non-parametric equations and PCA. I highly recommend taking this course to start your journey into the program.

Overall it is a very useful course. By the end of the course you should at least be aware of most of the popular models out there and know how and when to apply them. This course is NOT about teaching you how actually write code or use packages for building models. I think Dr. Sokol is clear about that, but many people seem to forget.

The homeworks are a bit annoying. I am not a fan of the peer review process, but there's not much else you can do with so many students in a course. Some of the TAs are extremely helpful, but I found some of the more active ones to be a bit rude. I ended up having to skim through the office hours every week in order to do the homework. I spent about 2 hours watching lectures/taking notes each week and ~4 hours on the homework.

The tests are fair, but are not like the homework. They test theory rather than practical knowledge - when would you use this model? What would adjusting this parameter do? Why would you get this result? Etc

I was proficient in R before this course. You could probably get by with just the basics though - complete the Swirl tutorial and make sure you understand how to manipulate data frames at a minimum.

Dr. Sokol is clearly a very intelligent guy and put together some interesting material for this course!

This was an excellent course to start the OMSS program. The lectures by Joel sokol were brilliant and the TAs were very helpful. I wouldn't miss the office hours. The exams were designed to ensure that you really understood the concepts and you are able to apply on your job. Highly recommend this for people joining OMSA.

Tough at the beginning due to inexperience with R/programming; Eased out towards the end. Great to understand analytical models at a high level.

Excellent Course from Dr. Sokol - Thoroughly enjoyed it.

Pros:

• Instructor who cares about students and learning
• No need to learn formulaes, just understand the fundamentals
• Good Homeworks which allows you to understand the concepts well if you put in the effort
• Exams very very conceptual
• Good grading system for the exams (partial marks for most questions)

Cons:

• Exam wordings were vague in many places - Answers could go either way for a few questions which could go either way
• Very low grading curve for Homeworks - An amazing job would net you 16/100 points, an average effort 15/100
• Sets a high expectation for other courses which doesn't look like it will be able to match looking at OMS central /s

All in all, definitely a must-take course in the first semester, pairs well with CSE-6040 which is more programming oriented (and in python, as opposed to the R focus in this course)

Pros:

1. Dr. Sokel used plain language to explain those not so simple concepts. Just amazing.
2. I'm used to learning visually, but this class forced me to learn by ears. Great experience.
3. The TAs were excellent.

Cons:

1. The homework was a bit wishy washy. Whatever effort you put in, you got either 90 or 100. Considering the time spent, I feel I should have learned more in R. I would like to be trained like a technician at least at the beginning of this Master program, and later on shift onto the arts of analytics.
2. Don't like the final exam as much as the two mid terms. The wording was too vague for some questions.

If you're new to OMSA, I would highly recommend taking this course first because, as the course title suggests, an introduction to analytical modeling that helps you get your feet wet in various models you will come across in the program/life.

Pros: The lectures were enjoyable to watch and Dr.Sokol does a good job of explaining concepts. The exams are mostly conceptual (which was a good thing for me since I've never coded in R before). The TAs were really responsive and other students in the class were very helpful as well.

Con: The homeworks can be very intimidating if you are new to R. If you have time, I suggest brushing up on it. Don't worry if it seems intimidating the first few weeks, it will get easier. There are really no practice exams and the exams are not like the homework at all so I wish going forward Dr.Sokol will provide a full length practice exam.

Overall a very good introduction to the program!

Great first class to take in the OMSA program.This course gave me a better idea of which elective I want to take.

The first few homework assignments took me awhile to complete due to the steep ramp up in R, which I hadn't used before the class. Get to use ARENA and PULP and complete some case studies later on in the course. Exams were fair, testing your understanding and not asking you to regurgitate facts.

Operates as a good breadth class that catches a bit of almost every common type of analytics modeling technique, without going too deep into the fine details of theory and model diagnostics. I found the first few projects to be the most difficult, but once I familiarized with the process of how to read libraries, search the web, and ask questions on piazza, it became a cakewalk in the latter half. I don't have a super strong background in the field (only ever used regression prior), and most weeks I could finish the lectures and accompanying assignment in a single dedicated day at the library. Assignments are student graded, which may sound questionable at first, but I found that as long as I made sure to clearly but a little bit of extra effort on every assignment I consistently got high grades. Exams are all very fair with level of depth equivalent to what one experiences doing the homework, but I found the final project to be a bit phoned in (role playing someone else's large scale data science project). Overall a solid foundation course for the program.

Perfect intro class to the Analytics program. Generally quite shallow but very broad. First couple of weeks are a challenge if you have no experience with R, but otherwise it's not too tough.

Took this in first semester of OMSA. I came in with little to no R experience. The first few homework assignments were pretty rough because of this, but they gradually got easier as the semester went on. I thought the material was very interesting and engaging throughout.

The last few homework assignments do not involve coding-we were basically given case studies and asked to determine a broad course of action. These are what you make them. Since everything is peer graded, you can get by with less than stellar work on these especially. The project is the same format as these homework assignments.

The tests are all conceptual in nature. They are somewhat frustrating because some of the questions can be interpreted multiple ways, but overall they're fair.

Great class!

It was my first semester in the OMSA program. I took it along with CSE6040 and MGT6754. The learning curve was steep in the first few weeks because I hadn’t been in school for almost 20 years and didn’t know anything about R.

The course was excellent, and the professor did a great job in the videos, course design, and communication with the students. The TAs were very helpful, especially in those office hours, showing code examples and answering questions. The course was very organized.

The weekly homework assignments helped me understand the concepts and learn R packages. I liked the fact that the peer grading emphasized on learning, not grades. It gave me much freedom and encouraged me to explore beyond the basics. I spent time reading related materials and watching other online videos to learn more about the topics covered. I think the course was very effective in teaching students how to think and approach real world problems.

The two midterms and one final exams were challenging but not unreasonable. The questions were effective in testing my understanding although some were not covered in the class. ProctorTrack was a pain to use, especially with some of the drag-and-drop type of questions. My laptop’s screen was too small to see all parts of the question.

The individual project was an excellent idea but it wasn’t too different from the last few homework assignments, where discussion and exploration were more important than getting a right answer. I wish there were more opportunities for discussion among students and more feedback from the professor or TAs. But I understand it’s not practical given the number of students.

It’s definitely a good foundational course for the program, and it prepared me well for the other classes later.

Learned a ton from this course, got a B. I have some undergraduate level work in statistics, coding, and appraisal of evidence / experimental design. Only took 1 course as I'm extremely busy and this is my first semester.

Homework every week, I skipped 3 (you get 2 free skips). Consider actually skipping the easy ones(not the first half of the semester), since they contribute less to learning. For example, without a background in any production level statistical learning it was the homeworks that gave me the necessary intuition and motivation to learn. The homework schedule feels relentless. I recommend starting the day it is released and committing no less than 5 hours to it. The first two weeks I was committing 20+ hours a week to learning R studio and R programming language. Then there was a week I had to learn linear algebra to really get what was going on in the homework, also a 20+ hour week. And the week we used pulp for optimization also a 20 hour week. This was probably not necessary if you are familiar with these subjects, but I was going in naïve.

The videos are short but contain a lot of information, watch them at least 3 times each. Piazza and Slack were both great for formal and informal discussions respectively There is an online book "Introduction to Statistical Learning with R". Check it out early. Also one of the first things you should do before you start the first homework is the R studio 'swirl' library. It will interactively teach you the basics of R programming language. https://otexts.org/fpp2/holt-winters.html This link will make all your time series work easier. Stack Overflow and Google constantly.

After the first test, it's a good idea to study optimization and objective functions in depth. Speaking of tests, the subject matter is what you need to pass it minimally, but how to use that knowledge to compare models and choose better or worse models in a given situation is what lets you get higher grades.

Set up proctor track a few times before you take your first test, and actually go through the process of set up, practice test, then FINISHING in the correct proctor track style. Submit each test question when you are absolutely sure you are done with it, and double check when you're done that all questions are submitted.

I watched the first many Office Hours in which they hold your hand and answer questions about coding, and lots of hints are given for the homework. I hope this doesn't change, in fact I think that giving the code approach away in office hours is what made me learn so much in the class because I built a fluency of the process allowing me to focus on making the right high level decisions with models and data instead of focusing on using the right function.

My cohort was awesome, I hope yours is as good.

Overall this course is a good introduction to analytics. You get exposure to a number of different models and the general concepts and applications of analytics.

The lectures are very helpful and provide a lot of material in a condensed series of videos. Exams are based mostly off the lectures rather than homeworks (only the final exam had any questions relating to the homework). Homeworks are all in R/Python/Area, with some additional explanation-type questions as well. Exams have zero coding questions and are mostly theoretical/conceptual.

A fantastic class to give a high-level overview into analytical models in use today. Dr. Sokol has taken the MOOC format and successfully boiled the concepts down into a presentable, easily digested way. Lectures are not particularly long (longest week was about 30 minutes of lecture), and introduce concepts in what is relate-able manner so that anyone could understand. Once finished with the topic example, the week is finished with an assignment in R. Note that the difficulty in doing these assignments comes out of unfamiliarity with R, and not because the homework itself is difficult.

Homeworks are designed to be used for analysis and towards the end are typically achievable with one or two lines of code for a total of 16% (~13-15 homework assignments, dropping the two lowest grades). As the semester progresses, the homeworks will take from 10 hours down to 1-2 hours per week. This is followed by a project (5-pageish paper), worth 9%, towards the end of the semester in the form of a high level report to discuss how one would (conceptually) use the models learned in class to solve a real-world problem.

Three exams (two midterms and one "final") are administered through ProctorTrack, each worth 25%. The final is cumulative with a focus on the final few weeks' worth of material. None of them are terribly difficult, and are designed to test your analytical thinking rather than just rote memorization. Examples include providing a real-world problem and then providing inference statements to choose from. These are also not designed to trick you, and appropriate credit will be awarded in the event of ambiguity.

Overall, a wonderful first class to take to get started on the OMSA program.

Strongly recommend 6501 course as a first course. Course delivers very useful knowledge that can use for real-world. To get A grade, I recommend focusing on assignments. Even though it is only 14% in grade, it tells everything that we need to know.

A great overview of data modeling in analytics. It is an overall survey through the data, so the lectures were quite easy to comprehend and quite clear. However, actually thinking in an analytic way is quite different - it was quite exciting to think of all the possibilities in which we could actually model the data, and get my hand on it. There was a slight confusion with the first exam which I, along with quite a few of others, completely botched and needed to work hard to save myself from the first mistake. Except for that, excellent TA, excellent course, excellent lectures, an absolute recommendation for the first course.

1. This course offers a broad overview of the different models and techniques used in analytics.
2. Most homeworks require programming in R so it helps learning it ahead of taking the class. Knowing R markdown would help you to write the homework report relatively fast, blending your code and comments in a nice HTML or PDF document.
3. If you take CSE 6040 concurrently or can program in Python, the optimization and simulation homeworks could be very easy.
4. The lectures do not cover the material in depth. Additional readings are provided only for the first modules in the course. Afterwards, you need to rely on your classmates suggestions in Piazza or Slack. Sometimes it may feel like the blind leading the blind :-), and you may get the same feeling from time to time when reading the comments on the peer-graded homeworks.
5. Useful references: for the first half of the course focused on classification and regression models, get James et al "Introduction to Statistical Learning with Applications in R" (free copy and codes available from the authors); for forecasting, check the online book by Hyndman at https://otexts.org/fpp2/; for probability models, download a probability model book from the library, like S. Ross, "Introduction to Probability Models" or Haigh, "Probabiilty models"; on optimization, any introductory management science or operations research textbook could help. Also check this other textbook, Shmueli, "Data Mining for Business Analytics: Concepts, Techniques, and Applications in R".

I really loved this class and learned a ton! One suggestion I'd make is to continue the additional reading recommendations throughout the course. The first few weeks had links for additional reading but they fell off after awhile. Since the lectures aren't very technical having reading recommendations where we can learn more of the details really helps.

Overall

I really enjoyed the course. It gives you a broad overview of several different analytical approaches and refinement methods. The content was interesting and Professor Sokol does a very nice job with the videos. Piazza and Slack communities are active and helpful. Would definitely recommend this as a first-semester course in the OMSA program.

Prepping for the course

• Learn R (you don't have to be a master, but proficiency would be very helpful).
• Learn either R Markdown or Jupyter Notebooks

Exams

• There are two Mid Term exams and a Final, which is cumulative.
• Each exam counts for 25% of your grade

The Good

• Course videos are very well done
• You get a high-level overview of several different analytical methods
• The homeworks give you a better idea of how the concepts are applied in a practical sense
• The last few homeworks and course project are more of a case-study, which lets you think critically about a multi-part problem.
• Exams aren't super time-consuming. Mid Terms were 1-1.5 hours and the final was 2-2.5 hours.

The Bad

• The exam questions can have tricky-wordings. You need to read the questions carefully (and then re-read them)!
• Peer grading on the homeworks and course project can be hit-or-miss. Luckily the TAs are available for regrades if you think you've been wronged.
• ProctorTrack
• There are a couple times in the semester where you have a lot going on between working on your homework, studying for a midterm, and doing peer reviews. Manage your time wisely. You can't really work-ahead too much in this course.

The Ugly

• There are 14 total Homeworks (lowest 2 get dropped). While these homeworks are helpful in learning the material, they can take a lot of time; particularly if you are striving for 100% (instead of just 90%). Despite the big time-commitment, the homeworks collectively count for only 16% of your grade, which means each one is just over 1% of your total grade.

This is a great first course for the OMSA. The material covered was broader than it was deep and provided an overview of a lot of analytics modeling concepts. The weekly homework assignments were reasonable in terms of difficulty and the time it takes to complete them. The homework provided an opportunity to implement the lecture material in practical programming exercises. The exams did not touch much on the programming aspect of the homework assignments but covered the same lecture material. The exams were fair. I would recommend that you take advantage of the practice exam before the first midterm because it gives you an idea of the type of questions that will be on all future exams.

It's a good first course for someone with a basic understanding of statistical concepts. Unfortunately for me, I do quite a fair bit of analytics and found the content underwhelming and assignments to be a chore, rather than actually inspiring me to be interested. To be impartial, I'll list 3 pros and 3 cons of the course as follows:

Pros:

1. Good introductory course - Pretty much how I would like to dip my toes into the water. It provides a broad overview of the concepts and tries to get candidates to apply them.
2. Easy workload - I listened to lectures on my way to work and averaged 1-1.5 hours on weekly assignments. Workload was manageable.
3. Good exam philosophy - Exams were focused on testing the understanding of concepts. This was a truly great experience and I commend Joel for the hard work. I'd imagine setting 3 fair and substantial exams to be quite a challenging task.

Cons:

1. Waste of time - I do a number of statistical modelling at work and felt that the materials were too shallow. Assignments were more annoying than constructive and I would have preferred to go into much more depth.
2. Can't think of any complains
3. Can't think of any complains

Overall a good introductory course, but I would have loved to go into the deep end.

This course is great as it gives you a great overview of what the OMSA program can offer you. Each week you cover one or two topics like machine learning algorithms, regression, missing data handling, and optimization. Then, you have to do homework using R (Python sometimes). This is where you'll spend most of your time if you want to get 100% on each homework.

BTW, the homework is peer graded. At first, i was a bit anxious about this process, but it turned out OK. I received only one grade that I felt was unfair, so i contacted the TAs about the situation and i got my homework regraded accordingly.

The exams were interesting, and you really need to have a strong understanding of the lectures' material. Make sure you revised math concepts before taking the exams (ANOVA, confidence intervals, what is a convex function...), as the teacher expects you to know them beforehand. Also, make sure you did the practice quiz before the first midterm, as it will help you understand how the ProctorTrack software works.

To sum up, take this course during your first semester. It will help you discover what subjects you will want to explore with the elective courses.

Pros:

• Overall, I think this course is a good introduction to the program. It gave a high level overview of many different analytics models that are covered in more depth through other classes in the program. Gives you an idea about potential classes you may like to take.
• Homework is practical, about simple applications.

•   TA sessions walk you through homework if you aren't sure what to do. 2 can be dropped. They took me about 4-5 hours each to get 100 on almost all (I was comfortable with R and python).
  

• Lectures are done well but do not go into much detail
• Get 9 day window to take tests, making it very convenient.

Cons:

• I understand the need for peer grading Homework and Projects since it is a MOOC, but the student experience unfortunately suffers from it. I never felt like the feedback was helpful to understand if my methods were used correctly. Comments were mostly “good job” or “did what was asked...”
• I think the tests were fair, but I found a few questions tricky in terms of the wording. They were multiple choice ("select one" & "select all" options).

Comments: Math of models is out of scope for this course.

Joel does a great job with this introductory class. Great lectures and relevant material that you can apply in the real world on day 1. The homeworks are tough, but you do feel like they are valuable. Exams are tough as well but not pure memorization; they focus on how you apply what you learned in real life. Overall, I think Joel nails this one and I believe that other professors should follow his lead on how to design an online class.

Overall, I think this is a decent course. I like that it is a broad overview of many different techniques in analytics. If you have an appetite for a deep understanding of a particular topic, you won't be satisfied by this course. If you understand that up front, then it should be fine.

That being said, I felt that some of the quizzes were way more in-depth than the material provided. I'm not sure how they expect someone to think very deeply about certain topics by watching a 5 minute video on it. The level of nuance that is expected to parse some of the test questions is insane to me. The homework assignments don't seem to align at all with the tested material (except for one question on the final exam which asked you to specify R functions you would use in certain situations). They also pretty much give you all the code anyway, so I'm not sure how useful the homework assignments are in the first place. The peer review process is hit or miss.

The range of quality of TA responses is also all over the charts from just giving you a link and telling you to further research it or providing a detailed response.

The TAs were flexible with assignments -- I knew I would be out of town one week far in to the course, so I planned ahead with them in order to be able to submit the homework on time.

And no, I didn't get a bad grade in the course, so I'm not just complaining because I did not do well.

Overall I enjoyed this course -- great overview of things to come in the OMSA program.

Would recommend taking either this or CSE 6040 as a first course.

This is my first semester at Georgia Tech and I enrolled in this course with CSE 6040 (Computing for Data Analysis). This course primarily uses R although it dabbles in other software. Tests are largely multiple choice answer based, which makes them fast to take but unforgiving towards mistakes. Joel is a great lecturer. He finds ways to add humors to most subjects he speaks on.

Introduction to Analytics Modeling is the first course anyone should take for the OMSA program. It gives a high level overview of analytics methods your standard classifiers to design of experiments to optimization. The homeworks provide practical experience with R and I think the TAs provide a decent amount of hand holding for anyone new to modelling or programming in R. Homeworks are weekly and you can drop two of them

This would have been a good choice to take first. The programming isn't super intensive, but a wide range of analytics models are introduced and the lectures are pretty good. The TA videos/office hours are all over the place. some were pretty helpful, some were a colossal waste of time. most hw assignments were pretty reasonable to complete. There was only one that I was completely lost on. Peer grading is kind of pointless. I didn't get any meaningful feedback. it seemed like the only thing that came out of it was people complaining that they got docked too much. The rubric is somewhat confusing and I think people had pretty varied ideas about what constituted additional effort. I thought the exams were surprisingly fair and well done. If you have a good understanding of the models and what their limitations and appropriate uses are, you should do pretty well. case studies were ok, but it might have been a good idea to publish a range of good examples and poor examples. proctortrack is a pain. overall an interesting class.

The professor is very knowledgeable, the videos are very informative and the assignments are very interesting. There is an assignment each week which would keep you involved. I just would like for it to be more collaborative as that's the intent of assignments but that didn't happen.

I thought this course was very well done and really is a perfect introduction to many of the options in analytics. If you have decent R skills, then it will make your life easier. Python helps too later in the course for some of the homework. The only downside is that this course moves quickly, each week is a new model, if not multiple models. In the summer, the same homework is given, but you have less weeks, which means each homework has more problems than in the fall or spring. I'd imagine the roughly 12 hours a week I spent on this course would be around 10 or less during the spring/fall.

Very high level overview of different models used in analytics. The course is not as technical as you would be led to believe. The R coding homeworks are very rudimentary and the TAs cover what you need to know for the coding homeworks. The mid-terms and exam are not technical/quantitative at all. They are completely conceptual and require no live coding. Overall the course was good as an introduction into widely used models and basic concepts in data science, but is not very technically difficult at all.

This was my first course in OMSA, and it is a great intro to the program. If you don't know R, then do attend all the TA sessions in the evenings to be able to complete the HWs (though you do have the option to drop 3 lowest HW scores). There is one HW where you use Arena for simulation, and 1 using SimPy for Python, but nothing too crazy. The exams focus more on concepts.

I enjoyed this course a lot. It is a whirlwind review of analytic methods. Its a mile wide and an inch deep. Expect to come out of it able to understand and use the language of analytics, but not with enough skill to really implement anything well., Assignments can be tricky, not a lot of programming guidance is given unless you watch the programming office hours(which I recommend). The tests are multiple choice style. I think the challenge of the tests is pretty fair vs the rest of the material. I only have two critical points, the case study part of the class should be more practically oriented(actually doing the problems) and the homeworks don't have enough weight on the grade(they take a ton of time and only count for 9%).

Good first course to take. Shallow overview of many topics, wish there was more focus on the actual implementation of models than with the conceptual understanding, but it IS an intro course.

An excellent introduction to Analytics. The course was designed to put you in the right mindset for identifying and determining how to use analytical models to solve real problems. R was used throughout, you should get the basics down early.

This course is a great overview of topics in analytics. It's very broad and touches on topics just enough that you're familiar with them. It gives you a taste of many topics to help you identify what you may have a passion for in the future. Dr. Sokol is a fantastic lecturer, and the videos are highly polished, well composed, and a breeze to understand. This was a great introduction to the program, and I wish all the courses were this well put together!

This course introduces a variety of analytics methods, and some software packages to implement them. Most of the weeks have an assignment to solve using an R package, a couple with Python puLP and simpy (or Arena simulator). Though one week I used Matlab, because I felt more comfortable. 3 weeks toward the end have case studies to write about (no programming). For the project, you choose a case study, more in-depth. Discussion with other students via Piazza or Slack is encouraged, but assignments are all individually completed. TA Office hrs are available to help with assignments as well. There is an assignment due almost every week. I had programming experience, but neither R nor Python. I recommed DataCamp.com to help learn those programming languages. It is free to try. The paid version is normally $300 per year, but after you complete 1 full lesson, they send an offer for 40% off.

Love this course very much, it introduced me to whole new world of analytics comprehensively, it will touch all concepts just enough to get familiar with and understand the model and be able to choose one. Programming in R and python is required in this course. Office hours are very helpful, in fact you will need it to be able to solve problems in weekly assignments (yeah assignments are weekly, get ready to have stressful weekend :) ).

Exams are multiple choice online proctored, it will test your understanding of all concepts you learnt and you need to reason which model to use in different case study during exams.

There is one course project which you have to analyze a real case (no real coding required), just a report.

The first bit of this course is intensive and comprehensive. Sometimes I have to stay up late to finish off the weekly programming homework. It introduced many analytics models I've never heard of. Thanks to this course, I have better idea of what courses to select next. The last bit of this course (case studies) is a bit loose. I think case studies could definitely be organised better. I think a large, research-like analytics focused project would be more suitable than report type homeworks.

I would recommend to take this class as a first course. For the first few weeks, homework can be time consuming to get a good grade if you are not familiar with R. Sometimes lecture materials were not enough to get 100 in homework as a grade scope of 100 required "a deeper solution than expected". TA's weekly office hour was very very helpful to clarify confusions in a homework. Few troll in peers can give you a low grade without any comment, which is frustrating. But, most of reviews I got was reasonable and helped me to improve. If they don't change the system in the next semester, your assignments will be reviewed by three different randomly chosen peers and assigned grade will be a median of the three grades.

Exams were fair. There were few tricky questions that require some thinking. What I liked the most for the exams was that it was designed to test your understanding on the models, not memorization.

Overall it is very good class and I've learned a lot through this class. I took CSE6040 and ISYE6501 at the same time. I recommend to take ISYE 6501 prior to CSE 6040 if you are considering one class among these two.

My background: Mechanical engineer with minimal programming experience. I did take GTs introduction to Computing which utilizes Python. However, this class uses R so take that for what it's worth.

Lectures: Lectures are well organized, sufficient material to have a clear understanding of the topic but not so much that you begin to zone out and lose interest. When studying for the final, I was able to review and watch all lectures within a few hours (2x speed).

Homework: Like many reviews, I will agree that the homework assignments can be disproportionate in terms of overall grade weight and time input, especially if you're like me with very little R programming experience coming into the class. Still, good practice utilizing the concepts learned in lectures. Peer grading can be very inconsistent, but generally forgiving overall.

Project: No coding, open ended. The opposite of HW. Equal to over half of HW grade weight but equal to the amount of work for a single HW. Some adjustments could be made between the two.

Quizzes (midterms): These can be pretty tricky so you really have to have a solid understanding of the material. An item that was only mentioned for a few seconds on a single slide could be a question on any quiz. A "cheat sheet" is allowed for all quizzes but I rarely used them. I believe they were a teaching mechanism to re-enforce learning more than they were as a reference on a quiz.

Professor/TAs: Solid. Dr. Sokol's personality really shows in the lectures and he clearly enjoys the material which also gets the student naturally engaged in learning. The TAs were very responsive in Piazza/office hours and also the slack group for the class so you never felt like you were on an island.

Tip: Join the slack group

Overall, excellent course.

I absolutely loved this course, and I'm very glad I chose it for my first semester. Dr. Sokol's lectures were engaging, informative, and clearly understandable. I don't know if there's much he could have done better with the lectures, they were that good. Even when he was working through formulae and variables, I was still able to follow along most of the time, and if I wasn't, just a couple repeat viewings and I was on the same page. The course surveys countless topics at a rapid speed, but I never felt like it was moving too fast. I'm amazed at how much I know now compared to the start of the semester.

Homeworks mostly use R for building and evaluating models. I'd had no experience with R before starting the course, so the early going was a bit rough, but once I got up to speed I was much more comfortable. The assignments are designed to set you on your own path of discovery and learning, and I found that I benefited the most from them when I approached them that way. On the weeks where I was more time-constrained or felt a bit lost in R, I found the TA office hours sessions to be very helpful. I never attended one live, I only ever watched video recordings. I did my best not to carbon copy everything the TAs had done, because they basically provided you with the answers most of the time - even something as simple as changing the names of variables helped me internalize the solution better than copying everything to the letter.

Peer reviewing has pros and cons, the biggest pro is that I saw a few really impressive submissions that actually taught me a thing or two (especially when people took the time to create Jupyter notebooks out of their submissions) but the worst con was reviewing some of the poorer submissions or the ones that just lifted the answers from the TA office hours with no elaboration. So depending on the luck of the draw you could find the peer review process to be educational or not. Also, feedback from other students was very inconsistent, sometimes you'd get great suggestions about how you could have done things differently, but most of the time it seems I got no feedback at all or something brief like "good job" with no detail.

Exams were very fair, and I didn't mind the proctoring as much as I thought I would (but I'm also not a parent so a quiet, private session was easy for me to manage). Exam questions usually required you to think through every answer and choose carefully, and I thought they did a good job testing my understanding of the material.

Overall I cannot recommend this course highly enough, and I strongly suggest taking it in your first semester as an OMSA student if you can.

Great course. I wish I would have taken this my first semester instead of my second. It is a great intro to Analytics as a whole. Dr. Sokol is an effective communicator in the lectures. Most of the interaction on the forums is with TAs. Dr. Sokol only made a few appearances on there. I highly recommend joining the Slack to supplement learning. Many students complained about the peer grading. I had no issue with the peer grading setup. Just make sure you do work that you are proud of, and you'll get good reviews.

Great Intro/Survey class for all of the ISYE classes. Joel is a great instructor who creates problems that force you to think about why you are answer something a particular way instead of just memorization.

Most of the other reviews have taken care of a lot of the details, so I'll keep it brief:

Good:

• Professor: Joel Sokol is a great teacher and made all the concepts very easy. He is great at putting things in relatable examples, and he has an obvious passion for the topics he teaches.
• Tests: I thought they were a good balance of difficulty.

OK:

• Peer Grading: I know and understand that peer grading is supposed to help you by helping you see problems solved in different ways, and because of that I didn't put this in the "bad" section. But I still hate doing it, so I didn't put it in the "good" section either.

Bad:

• Can't work ahead: For reasons that are pretty obvious, I want to be able to work ahead so that if I have a particularly busy week at work, or go out of town, or the like, I can work ahead and be OK. There were two homeowners I couldn't do because you have to do them in a 3 day window and I was out of town.

My Background:

• 22 years old, soon to be married, graduated undergrad in May 2017.
• Coding Experience: Minimal (light experience with non-Python coding languages)
• Statistics Experience: Moderate (4 undergrad level stat courses)
• Math Experience: Moderate (peaked at 2nd year Calculus)

Good pairing with CSE 6040 The two courses nested nicely. Early on, learning R and Python simultaneously was a bit hectic, but as ISYE 6501 used Python pulp for an optimization homework, and I was able to merge both learnings using reticulate and rpy2. Also, as 6501 backed off the coding and entered the "case study" phase, 6040 entered the "application to models" phase, implementing models that I learned in the beginning of 6501.

**[Sean Connery Accent] One Submit Only [/Sean Connery Accent] ** All homeworks, the project, and the "quizzes" were "single submit." This meant a lot of double/triple checking prior to clicking.

No real working ahead Everything was pretty much on a weekly schedule, which meant you were ALWAYS thinking of AT LEAST 2 topics at a time (and more with exam reviews). Very little opportunity to work ahead, and if you have something going on during the week or over the weekend, forget about it):

• Monday 0: Release video lectures for Topic 0
• Thursday 0: Release homework for Topic 0
• Monday 1: Release video lectures for Topic 1
• Thursday 1: Release homework for Topic 1
• Thursday 1: Homework for Topic 0 due
• Thursday 1: Homework for Topic 0 open to peer reviews
• Sunday 1: Peer reviews for Topic 0 due
• Monday 2: Release videos for Topic 2
• Thursday 2: Release homework for Topic 2
• Thursday 2: Homework for topic 1 due
• Thursday 2: Release homework for topic 1 for peer review
• Sunday 2: Peer reviews for topic 1 due
• Monday 3: Release video lectures for Topic 3
• Lather, Rinse, Repeat until end of term, not forgetting the exams (they called them "quizzes" but they're 75% of the grade for 3 events, so...)

Grades != Time Commited

• Each HW was ~1% of the entire course grade, but could be several hours of work in an open-ended discussion-type answer.
• The Course Project was ~9%, but slightly less work than a homework grade.
• Each of the 2 midterms were multiple choice and could be completed in less than 30 minutes, and the final was twice as long, but still less than an hour, but all told were 75% of the grade.

ISYE 6501 Slack Was the primary means of sanity and mutual support. Much better than the Piazza mess.

Overall: Great class - it's so applicable to the real world that I would frequently learn something in class and hear about it within my job within a couple weeks. (I am a new member of a data science team). Dr. Sokol structures the lectures well to be informative and efficient. Bonus: dad jokes.

Homework: Challenging at first without R experience, but you quickly learn how to handle the language and to debug problems that pop up - plus, the TA hours and ability to work with peers helps tremendously. I am a strong believer in Dr. Sokol's statement that we will continually need to learn new programming languages and packages in our career, and so being thrown into things without being hand-held is vaulable and realistic.

Project - Fairly simple, though peer grading makes it a bit scary! NOT a group project, which I appreciated.

Exams - While I received decent grades, I will say the exams were tricky (mostly #2 and #3). Dr. Sokol knows how to write an exam to make sure you REALLY know the information more than surface-level. Some questions are easy to interpret incorrectly, and that is my only gripe with the class.

Prereqs - While there are technically prereqs for probability, linear algebra, calculus, and programming experience, the primary one that you need is some programming experience. Coming in completely blind to programming would have been a significant handicap. Probability, linear algebra, and calculus are fairly minor (though not COMPLETELY invisible - eigenvectors and the calculation of linear, convex quadratic, general convex, and non-convex equations all do come up).

Overall - definitely recommend this class, even as a standalone for non-degree students.

Good course to start learning about Analytics models. Sets the good foundation for R, data models. Knowing R ahead of the course is plus but not must. You catch up in the initial 3-4 home works with extra study. Good TA assistance on assignments using R. Do brush up your knowledge about linear algebra and basic stats for a better understanding. Homeworks take a long time and only grades 16%. Peer review many times is subjective and can't respond to reviews unless you have been graded pathetically in which scenario you have to go via TAs for re-evaluation. 3 peer reviews on each assignment. Some people are just mean in the review processing. You will learn from other a lot during reviewing other's HWs. 3 midterms are fair but tricky on wordings and make you overthink. Questions are graded with high weight and a wrong answer can cost you equivalent to 4 HWs. Make sure you understand concepts and applications to get good marks in midterms. The project is easy but again subjective nature of peer review is something can take out your points for no reason. Rubrix for the project is not well defined. Overall contents of videos are good. Make your own notes and be active on Piazza and slack. TA's are doing a very good job and always got the answer to questions within few hours. Even Dr Sokol answers and guides when needed, which is really great.

Overall great course, you learn a lot.

I usually used to spend 30 min every day (give and take) in reading material for the week and spending on HW during weekdays and weekends more like 10(+) hours in self and deep study and homework. Some homework used to be longer and can take more hours, so in general demands on an avg 20 hrs each week.

Yep! everything is in R. However, very interesting course, it introduces you to various analytical methods. Prof. Sokol is really fun, succinct and engaging.

Amazing course as an intro to OMSA or for anyone willing to get a broad understanding of Analytics concepts and techniques with R language.

• Content-wise, I have found this course very condensed, there is definitely a lot to learn. The online format of the course is very succinct therefore additional readings definitely help in building the intuition. To do so, I have relied on various books, sites and blogs, but also used the discussion forums which are definitely a key to success for this course. This course is greatly designed and the knowledge acquired after finishing this course is definitely its strength.
• Logistics-wise, on average one homework per week is to be submitted, the difficulty varies based on the content of the course, I have spend 8 to 10 hours a week between the course and the homework. The peer-grading can be frustrating at times, but equally rewarding when receiving a feedback on the submitted homework, which is not necessary the case. The weight of the homework in the overall grade is quite low compared to the level of effort. However, the time spent on the homeworks is useful to prepare for the midterms and the exams, which are more concept-based and of average difficulty.
• Prerequisites-wise, elementary knowledge of R is definitely helpful to focus on the concepts rather than wasting time learning the language in parallel. Dedicated office hours will also help with R if required. Basic knowledge of math is required but the course is not deep diving into complex math formulas and theorems, which made it easy from that perspective.
• Course team-wise, the course is provided in the perfect format, appropriate for online learning but still ensuring the quality of traditional academics. Real-life or project examples are shared by Dr Sokol with a pinch of humor throughout the course, which makes it very captivating and interesting as well, beyond the theory. Last but not least, the TAs are VERY present and supportive, which contributed a lot to the success of the cohort.

A mile deep and inch short survey type program. Does a good job of covering the gist of many analytic topics. Homeworks could be challenging especially if you don't have experience in R. Even so the TAs office hours often spoon fed solutions. Tests were difficult but not unreasonable. The final was easier than the two midterms. Overall a very good introduction to analytic modeling and the OMSA program.

Nothing easy about this class, occasionally very hard (hello SVM equations week 2). But the best class I have ever taken. Someone here said "excellent pacing and length" - totally agree with that. And Dr spool is a very engaging lecturer to boot. Recommend doing an intro to R to reduce workload in class. I got the equivalent of an A but the workload was especially intense the first 4-5 weeks as I had to cope with both the material and learning to use R. Admittedly I didn't use the office hours much, when I did tune in I found a lot of dead time between interesting bits and I lost patience. Stack overflow (and occasionally RTFM) to the rescue... But to recap how I started, it's a fantastic course.

Very well-done course. It's hard to imagine a better introduction to the Analytics program -- you should definitely take this your first semester. This course does a great job of surveying the world of what most people are calling "data science" nowadays -- basic data preparation plus a wide array of both machine learning and traditional statistical modeling techniques. If you put a good amount of time and effort into the homework (and don't just copy the TA office-hour solutions), this class also goes surprisingly deep into some topics. You WILL want some experience with R prior to the course or you will be spending double the hours/week. It is also a great idea to get familiar with RMarkdown documents since they are a great format for submitting homework.

Although I ended up with an A, my only complaints about the class relate to the grade weighting and some exam questions. First, the homework and course project are low weight compared to the amount of work required and understanding gained (16% hw, 9% project). In addition, one feels a bit cheated when the course project is merely peer-graded and your peers don't provide much feedback. The peer grading in general was not very helpful as I rarely got any comments, just a numeric score -- I would learn more with tougher HW feedback from an instructor. The exams felt unnecessarily high-pressure and high-stakes, because they were (75% of grade for 3 exams). While about 80% of the exam questions were pretty easy, 5-10% were subtle/difficult, and about 10-15% were just badly constructed -- either expecting us to "think on our feet" about applying/combining concepts that only got minimal lecture time, or having ambiguous wording/intent, or relying on a very artificial set-up, or making too many assumptions and providing too little context. Some were major traps for people prone to overthinking. In addition, the exams leaned toward fewer questions with higher point values each, which can really bite you. Fortunately there's a fairly generous curve that will add 5-6 % points. In general, if you want a solid pre-curve A, you should do extra reading on each modeling technique covered and your studying should focus more on big-picture, situational application and pros/cons of different techniques.

All in all though, I thoroughly enjoyed this course and learned a ton. Dr. Sokol's lecture videos are engaging and clear, including many examples from his own consulting experience. It's obvious he loves the subject too.

I found the tests in this course very challenging, and the homeworks required way more time than they were worth as a grade percentage. I learned a great deal and the amount of material covered is quite broad.

Great class! The videos with Dr. Sokol are very engaging and have excellent pacing and length. The assessment methods for the class keep you on pace and are a good reflection of the material (16% weekly homework, 9% project, 25% midterm 1, 25% midterm 2, 25% final exam). Your incoming level of R knowledge will greatly impact your hours/wk workload. Be sure to checkout the weekly TA office hours if you're new to R. The course provided a great introduction to analytics and had the right amount of depth and breadth to spark excitement and curiosity for the material.

This class was a whirlwind - you cover SO MUCH GROUND! As other reviewers said, it's a great intro to just about every Analytics concept that you're cover to see as part of the OMSA program. And, for me at least, it helped me figure out which electives I wanted to take.

Now, with that said, it's a challenging class especially if you've never used R before. I find it to be a pretty finicky programming language, and there were nights where I struggled for an hour or two to figure out what I was doing. By the time I was done, I felt like I had learned a good bit, though, so maybe it was worth it.

Homeworks aren't bad, though a bit time consuming. Midterms/final were a little bit tricky - you really have to sit and think through a question and pick the answer that seems the most right. You get a handwritten cheat sheet for each midterm and 2 of them for the final - and these saved me. But you need to put a lot on them since briefly covered concepts could be on the exams.

Project for the class was simple enough and followed the format of other case studies.

It's not easy, but you learn a whole lot, and it's a great introduction to what to expect for the rest of the OMSA program.

This is essentially a sample platter of the analytics program. You'll get a little regression, operations research, time series analysis, etc. Now, you'd think, given the survey nature of the class that the topics would be fairly high level, but they weren't. For regression, for example, we got into things like elastic net, and regularization, etc. Tests were not too bad if you did a reasonable amount of preparation (you were allowed a cheat sheet, both sides). Interestingly, I found that my cheat sheets really didn't help. I found that for most questions, I either knew the answers, or nothing in my notes would have mattered anyway. If you come into this class with no R experience, get ready for a steep learning curve. In my case, I certainly had some basic knowledge, but found that having a copy of "Introduction to Statistical Learning" came in handy. At the beginning of the course many people freaked out due to the seemingly steep R learning curve. But, as was pointed out, the homeworks were designed more for learning than testing. As it was, homework counted for relatively little of the grade, and submitting ANYTHING would get you a 50% score. And, since homework only counted for 16% of the grade total, that means you could have done almost nothing and ace the test and come in a 92% course grade. Having said that, I really tried on each homework assignment to do the very best I could. After the early weeks, the latter part of the course was a bit easier, with the written project more conceptual. Overall, good course, certainly increased my confidence in figuring things out with R.

This class was a lot of fun! I thought it covered most of the basics and is a great way to start OMSA. You will find this course really easy if you are already familiar with R and modeling. Dr. Sokol does a great job teaching this class and is very good at explaining difficult concepts.

Great class. I encourage everyone to take this class as their first class in OMSA program. You'll know if this degree program is for you by the end of this class. I came into this class with programming knowledge (python) and some linear algebra but almost zero R experience and no analytics experience. This class was more time consuming for me because I've had very little exposure to analytical modelling. If you come into this class with experience how much time you need each week will go down considerably. Early on in the course getting up to speed on R was a challenge but once I was comfortable in R the amount of time required for the homework goes down quite a bit. There is homework almost every week. It's well designed and if you actually try to learn you'll learn quite a bit. I encourage people to use the TA office hours as confirmation of approach after you have tried to solve the homework on your own. You'll learn far less if you just watch the TA office hours and copy their code. Exams are hard but fair. Dr Sokol does a good job of making you think. The videos for this class are very well done and to me shows Dr Sokol's strong understanding of the MOOC format. The videos are short and densely packed with information making them quick and easy to review while at the same time, packed with information.

This course covered a lot of material at a surface level, but it was a great intro course and it really primes you for all of the other course work that may come later.

ISYE 6501 gives a high-level overview of data preparation, modeling, and other analytical techniques (such as simulation). It's a great introduction to the OMS Analytics program. The course consists of: 3 exams (2 midterms and 1 final) each worth 25%, homework worth a collective 16%, and a final project worth 9%.

I came into this class not knowing any R, which is the language primarily used for the homework assignments. If you have any experience programming, you'll immediately notice corollaries to whatever language you are familiar with. The TA Office Hours are very useful for learning how to use R if you're struggling. The homework assignments help you understand how to use the concepts from the lectures in an applied way. The course project is essentially a slightly longer homework assignment, on a topic of your own choosing. It follows a basic case study format. The exams are multiple choice and drag & drop. They're tricky, but not overly difficult. Anything covered in the lectures is fair game -- even concepts that are only touched upon briefly.

This was a great way to start the OMSA program; it's basically a high-level overview of everything the program offers. I knew R coming in, so I found the assignments fairly easy once I figured out 1) what the assignment was actually looking for (Office Hours were great for this) and 2) what R function/library to use. That said, it is worth having some programming background and taking an R intro course before starting the class since there is a very steep learning curve. Lecture videos were interesting and easy to follow. The course project, despite being a fairly large portion of the grade, does not take a ton of time to complete. Peer grading could be a struggle since the grading standards are fairly open-ended. The exams are multiple choice and matching, but are actually quite tricky. None of the concepts are overly difficult to comprehend, but the question phrasing and deeper connections make the exams more challenging than you would expect. Weekly homework (due a week later), plus a course project (given a month to complete) and three equally-weighted exams spaced throughout the semester.