View on GitHub

Software Engineering Practices for Data Science

The habits that keep research code working: isolating dependencies, keeping secrets out of version control, using git and GitHub for real, and scaffolding new projects on a standard layout instead of an ad hoc folder — plus the argument for why any of this matters in the first place.

Ordered easiest to hardest.

1. Virtual Environments: Isolating Project Dependencies
2. Keeping Secrets Out of Git: .env & python-dotenv
3. Git & GitHub Desktop: Version Control from Scratch
4. Cookiecutter Data Science: A Standard Project Skeleton
5. Science as Amateur Software Development (talk)