Skip to content

Public release checklist

The repository is technically ready for public review after this documentation MR is merged.

Repository settings

The GitLab connector used during development cannot change project metadata or visibility, so complete these settings in GitLab after merge:

  • Set visibility to Public.
  • Set description to: Reproducible causal inference and econometrics: quasi-experiments, experiments, Causal ML, MMM, Bayesian models, diagnostics, and production workflows.
  • Add topics:
  • causal-inference
  • econometrics
  • difference-in-differences
  • causal-machine-learning
  • marketing-mix-modeling
  • bayesian-statistics
  • time-series
  • python
  • Verify GitLab Pages is enabled and the documentation URL resolves.
  • Consider disabling the wiki if repository documentation is the canonical source.

Release checks

  • Merge the v1.0.0 release-preparation MR.
  • Confirm the default-branch pipeline passes.
  • Confirm the Pages job publishes successfully.
  • Run the retail pricing case study from a clean environment.
  • Confirm poetry run mkdocs build --strict passes.
  • Tag the merged commit as v1.0.0.
  • Create the GitLab release using CHANGELOG.md.
  • Check that CITATION.cff is visible at repository root.
  • Verify the MIT license is detected by GitLab.
  • Verify README links after changing project visibility.

Portfolio review

A reviewer should be able to answer, without reading implementation code:

  1. What is the causal estimand?
  2. Which assumptions identify it?
  3. Which diagnostics probe those assumptions?
  4. How does the estimator behave when assumptions fail?
  5. How is uncertainty represented?
  6. How is a research estimator turned into a reproducible workflow?
  7. What is the business decision in the featured case study?

If any of those questions still requires reverse-engineering source code, the documentation is not finished.