This was my second class in the program. As a software engineer working on medical devices, the class met my expectations.
Tips
- Take the class if you want to learn about health data interoperability, want to take a class with no quizzes/tests, and/or want to work on a modern, resume-worthy project.
- I would not recommend taking a project with an external mentor. My team's mentor was especially flaky. It's simply not worth the "real world" experience. If you don't want a project with an external mentor, don't put any external mentors on your preference list.
- Extra-credit assignments generally aren't worthwhile (lack of direction & reward), but the peer reviews are pretty easy and can be pretty interesting for additional perspective.
Difficulty
I rated the class "easy" on the merits of "it's a guaranteed A if you do all the work". That said, the class is still a decent time commitment with lots of concurrent deadlines, which creates some mental fatigue.
Prerequisites
- JavaScript
- Java
- R (for extra credit only)
The team projects are all web apps, so you should have at least one team member with HTML/CSS/JS skills and at least one team member that can build a backend API service. I have strong web development experience and used the project as an opportunity to learn ReactJS.
Workload
The workload was poorly distributed. The first half of the class was overloaded with multiple individual assignments due weekly, while the last few weeks were nothing but the group project sprints. I estimate 12-15 hours per week average for the first half of the semester with one or two weeks easily at 20+ hours due to poor scheduling by the TAs. The writing assignments were underspecified with very loose length & detail guidelines, so it was easy to overwork many of the assignments.
The workload of the group project will depend entirely on your group. I took ownership as Project Manager and the most experienced frontend developer, so I spent at least 10 hours per week on the group project & weekly sprint deliverables.
Lectures
The Udacity lectures were marginally useful for a few labs. The live lectures from Dr. Duke were occasionally useful or interesting, especially in context of the ongoing COVID-19 pandemic, but completely optional.
Assignments
Major work consisted of:
- 6 Case Studies (short writing)
- 10 Labs (mix of long writing, short assignment, INGInious grading)
- Individual data analytics project (4 deliverables)
- Team Project (10x 1-week sprints, 4 video presentations)
- Reviews (course/peer/mentor reviews)
No quizzes/exams, which is a big plus.
The written assignments are underspecified, but graded generously. One annoying lab required jumping to a bunch of public healthcare websites, creating free accounts. The autograded INGInious assignments were moderately challenging, but had 20 attempts per item with no penalties.
Extra credit ("stretch assignment") opportunities:
- Peer reviews for writing assignments
- One case study
- RStudio work with the individual project
- Random points added to required assignments (above-and-beyond credit?)
The TA's won't grade any stretch assignments unless your final grade is below the 90% (A) cutoff and they didn't disclose how much these extra credit assignments were worth.
Individual Project
This project is a straightforward data analytics project with GT-hosted OHDSI Atlas software. The project isn't hard as long as your chosen hypothesis is viable with the given Medicare claims data. The live lectures step through the major tasks required for completing the project.
Team Project
This was a mixed bag for me. Many of my team members were simultaneously attending hard classes, so most team members weren't as invested or engaged in the project.
The project was chosen by ranking top-10 choices from a list of 44 potential projects. 15 were sponsored by external mentors. We got our rank #10.
I enjoyed working on a "real-world" project with modern technology (ReactJS, Spring, MongoDB, Docker, Kubernetes, Drone.io). This was easily the best feature of the class.