2025-03-01 · Amelia Costa

Notebook Hygiene for Pipeline Courses

Image unavailable

Jupyter notebooks tempt out-of-order execution. We require a top cell with pinned versions and a make export target that produces a reviewed HTML artifact. Mentors comment on the artifact, not a mystery kernel state. Students tag heavy cells with expected runtime on M1/M2 laptops so peers budget study time. That small note reduces dropout spikes in week four when assignments get chunky. We also ask for a data dictionary markdown file beside the notebook. Analysts downstream thank you, and it mirrors real job expectations more than a polished chart alone.

Tags: pandas, notebooks, data