Lesson 2 of 38 · Written learning
Set up a reproducible learning workspace
Prepare Python, notebooks, Git, and a repeatable place to run experiments.
Start with Python and a project folder. A virtual environment keeps one project's packages separate from another's. A notebook is useful for short experiments because code, output, and explanation can live together. A normal Python file is often better for reusable tools and larger applications.
Record the experiment, not just the result
Use Git to save named changes over time. A commit should describe a coherent change, such as “load and inspect the text file.” Record the Python version, important package versions, dataset source, and the command used to run the program. Reproducibility does not require fancy infrastructure; it requires careful notes and repeatable steps.
- Run a tiny Python cell and print a result before installing more tools.
- Use a CPU for early exercises; check a hosted notebook's current quotas and terms before relying on it.
- Keep passwords and API keys out of notebooks and Git repositories.
- When an error occurs, read the final lines of the traceback and reduce the problem to the smallest failing example.
Return to the Environment and prerequisites stage · Suggested timing is flexible; move at your own pace.