When teaching a course on biomedical data science, just the initial set up can trip up learning. What versions of R are students using? What if their computer doesn’t let them install a necessary package? How can they access the files they need, when they need them? If these questions aren’t easily answerable, students (and teachers) will need to spend time tinkering with their tools rather than applying them in the lesson.
Chirag Patel, Associate Professor of Biomedical Informatics at Harvard Medical School, faced this challenge when developing Harvard’s Data Science for Medical Decision Making course. The class focused on applying data science techniques to improve rational medical decision making.
Students needed access to large datasets, computing clusters, and R and Python environments with the goal of running high-quality biomedical analysis. Students also needed to be able to show and verify work — something critical for replicable medical findings. Normally, the process of figuring out logins, sending documents by email, and configuring workspaces would impose a huge workload on the course instructor at the beginning of the semester, just when they are busiest.
Professor Patel discovered that RStudio Cloud is a solution that supports his students through efficient onboarding and reproducible analysis.
Efficient onboarding for learning programming
Professor Patel realized that a critical first step in helping his class succeed is getting them started quickly. He wanted to find a way that students could focus more on learning to become better consumers of data rather than figuring out how to set up their laptops.
Professor Patel decided to select RStudio Cloud as a platform for his class. It allowed students to quickly dive into data by starting up a computing environment directly in their browser without installing any new software. Package installation and versioning were already taken care of, and Professor Patel could distribute his curriculum by copying his GitHub repositories directly into the workspace.
Reproducible analysis for better medical decision making
With installation out of the way, Professor Patel was able to focus on another key lesson of this course: code-first, reproducible analysis. Using RStudio Cloud, students shared their work in an environment that others could run. Their thought process was reflected in the code and others could review it step by step. Fellow classmates could run the analysis themselves and replicate the findings.
More than just being a useful teaching technique, code-first data science reinforced Professor Patel’s belief that analytic choices should be clear to the viewer so that they can make informed decisions based on the results.
Better experience, better data consumers
RStudio Cloud removed the barrier of setting up a computing environment while allowing for reproducible data analysis. Professor Patel’s class could instead focus on the data science approaches, interpretation, and results to become better, more active consumers of data. Read more about Professor Patel’s experience with RStudio Cloud on our Customer Stories page.
To learn more on teaching and learning with RStudio Cloud, please check out these resources: