In 2023, we use GitHub for selected parts of the assessment-related analyses. Specifically, the analyses where we would like to:
-
Track incremental development. GitHub commits describe which changes were made when, why, and by whom. While working on complex analyses such as stepwise model development, this can be helpful to clarify the logic, reduce cognitive load, catch accidental errors, and guide the next steps in the exploratory analysis.
-
Strengthen reproducibility. Compared to traditional Penguin directories, GitHub repositories tend to contain only the required files and be clearly organized, representing an analysis that is more likely to run on any machine. Any team member can clone a copy of the repository on their laptop and check if the analysis runs.
-
Facilitate collaboration. GitHub provides tools and workflows to help team members work together on a specific analysis.
-
Strengthen open science. We can switch a specific analysis from private to public at an appropriate time, inviting PAW and SC to examine how things were done.
-
Automatically backup. This means we can explore or restore any previous state and there is no danger of things being accidentally deleted.
-
Guarantee an exact identical archive (clone) on Penguin of the work that was done, while enjoying the speed and efficiency of working on the laptop hard drive.