Track

Python

Recreating an environment, building something installable, and making work that still runs in five years.

Python is easy to start and surprisingly deep to hand to someone else. This track is about that second part. Of more than a million public Jupyter notebooks, about four per cent reproduce their own recorded results — and the gap between an analysis that ran once and a repository a stranger can clone is mostly not a Python problem at all.

On Collegica

Python for Reproducible ResearchArticleFour per cent of published notebooks reproduce. Getting from one that works on your machine to a repository someone else can run, in eleven files. Environments Are Not PackagesArticleWhy pixi and Poetry are not competing tools, what a lockfile can and cannot buy you, and where uv lands. Publish Your First Python PackageGuideAn introduction to Poetry, from poetry new through to a package on PyPI — with a video walkthrough. Top Five Books for Learning PythonSlidesWhere to start, and what to read next.

Where to begin

The two articles are about why the pieces are arranged the way they are; the Poetry introduction shows you the commands.

  • Python for Reproducible Research — the arc from a notebook to a repository someone else can clone, with a worked example that was built and run before the article describing it was written
  • Environments Are Not Packages — an environment is a machine state you recreate; a package is an artefact you distribute, and almost every packaging frustration comes from confusing the two
  • What none of it buys you — reproducible is the floor, not the ceiling