0.5.0

API changes

Be aware that this release contains some breaking API changes, which should be confined only to the SyntheticControl experiment class. The docs and docstrings have been updated to reflect the new API. This was necessary to enable the main new feature of this release... we can now do synthetic control analyses with multiple treated units in a more coherent way.

As part of this, there was some behind the scenes refactoring work. All our tests pass and there should be zero visible change to users, but if there are any edge cases not caught by our tests then we welcome bug reports.

What's Changed

New Features 🎉

• Enable multiple treated units in synthetic control quasi experiments by @drbenvincent in #494

Documentation 📖

• Update difference in difference docs to show TWFE formulation by @drbenvincent in #482
• Add details about causal impact in the interrupted time series docs by @drbenvincent in #504
• Interrupted time series docs - more explanation and info on extracting causal impact measures by @drbenvincent in #501

Maintenance 🔧

• API change for the SyntheticControl experiment class by @drbenvincent in #460
• Remove docs/Makefile, update root Makefile for docs build, and update. by @Sohaibameer204 in #483
• Exclude causalpy/tests from code coverage by @drbenvincent in #493

New Contributors

• @williambdean made their first contribution in #489

Full Changelog: 0.4.2...0.5.0

0.4.2

What's Changed

Bugfixes 🐛

• Replaced MutableData with Data by @klaushajdaraj in #455
• fix bug: obs_indx -> obs_ind and improve PyMCModel._data_setter by @drbenvincent in #462

Documentation 📖

• Update badges on GitHub landing page by @drbenvincent in #465
• Add new list of written learning resources to the docs by @HPCurtis in #447
• Fix CI badge to properly reflect main branch status by @Nick1200000 in #476

Maintenance 🔧

• Separate InterruptedTimeSeries and SyntheticControl classes, update uml by @drbenvincent in #458
• Fix PyMC site reference by @emmanuel-ferdman in #466
• Auto-label pre-commit autoupdate PRs with "no releasenotes" by @Sohaibameer204 in #471

New Contributors

• @klaushajdaraj made their first contribution in #455
• @emmanuel-ferdman made their first contribution in #466
• @Sohaibameer204 made their first contribution in #471
• @Nick1200000 made their first contribution in #476

Full Changelog: 0.4.1...0.4.2

0.4.1

What's Changed

New Features 🎉

• Get plot data for prepostfit experiments by @lpoug in #438

Bugfixes 🐛

• Fix SyntaxWarning by making some docstrings raw by @drbenvincent in #414
• Calculate goodness of fit (r2_score) based on posterior expectation rather than posterior prediction by @drbenvincent in #429

Documentation 📖

• Enable ruff formatting of docstrings by @drbenvincent in #430
• Update docs to make clear sharp regression discontinuity design by @HPCurtis in #437
• Unsquash images in docs by @drbenvincent in #439

Maintenance 🔧

• Add test for API stability for a CausalPy example in Causal Inference and Discovery in Python by Aleksander Molak by @arthurmello in #407
• Implement the codespell pre-commit hook by @westford14 in #408
• Codespell CI Job by @westford14 in #411
• Sync dependencies in environment.yml and pyproject.toml by @drbenvincent in #415
• Update README and home page of the docs by @drbenvincent in #423
• Update head of apache and set start year to 2022 by @lucianopaz in #441
• New GitHub workflow to generate PR's which update the UML diagrams by @drbenvincent in #442
• fixes for uml github action by @drbenvincent in #446

New Contributors

• @arthurmello made their first contribution in #407
• @westford14 made their first contribution in #408
• @HPCurtis made their first contribution in #437
• @github-actions made their first contribution in #448
• @lpoug made their first contribution in #438

Full Changelog: 0.4.0...0.4.1

0.4.0

What's Changed

Major Changes 🛠

• Major code refactor to unify quasi experiment classes by @drbenvincent in #381

New Features 🎉

• Add example of multiple geo lift test analysis by @drbenvincent in #338

Documentation 📖

• Change docs theme by @drbenvincent in #371
• fix docstring example rendering by @OriolAbril in #376
• Add page of causal video resources from PyMC community into the docs by @ParthJohri in #379
• use labs theme from external repo by @OriolAbril in #385
• fix missing notebooks in docs by @drbenvincent in #398

New Contributors

• @ParthJohri made their first contribution in #379

Full Changelog: 0.3.1...0.4.0

0.3.1

What's Changed

New Features 🎉

• Add support for Python 3.12 by @drbenvincent in #347
• Enable summary method for all currently implemented frequentist experiments by @drbenvincent in #355
• Improve formatting of model coefficients in summary method in pymc-based experiments by @drbenvincent in #359

Documentation 📖

• Add docs on justifying instruments in the IV approach by @NathanielF in #345
• Various improvements to documentation infrastructure by @OriolAbril in #354

Maintenance 🔧

• Update codecov-action@v3 to v4 by @drbenvincent in #349
• Simplify remote GitHub actions by building test environment only once by @drbenvincent in #351
• Fix up ruff pre-commit checks to operate on notebooks by @drbenvincent in #352
• Remove all remaining traces of nbqa from the repo by @drbenvincent in #353

Full Changelog: 0.3.0...0.3.1

0.3.0

What's Changed

New Features 🎉

• Add propensity weighting schemes and covariate balance plot functionality by @NathanielF in #311

Bugfixes 🐛

• Fix FutureError warning, which requires pymc>=5.14.0 by @drbenvincent in #325
• Fix failing doctests by @drbenvincent in #330

Documentation 📖

• Add new knowledge base page on causal DAGS for quasi-experiments by @drbenvincent in #321
• Add Propensity Score DAG to Knowledge Base by @NathanielF in #337

Maintenance 🔧

• Revert the GitHub Actions update due to Node 16 sunset. by @anevolbap in #316
• Automate pre commit updates by @juanitorduz in #322
• Move to a conda only install of pymc (local dev workflow) by @drbenvincent in #304
• Fix #329: Add random_seed to the InstrumentalVariableRegression.fit method. by @anevolbap in #335

New Contributors

• @pre-commit-ci made their first contribution in #334

Full Changelog: 0.2.3...0.3.0

0.2.3

What's Changed

Bugfixes 🐛

• Fix #300: Set round_to as an optional argument. by @anevolbap in #301
• Fix #299: Rounding issue for DiD results. by @anevolbap in #302
• Fix #259: Annotation in did_skl notebook plot. by @anevolbap in #306

Documentation 📖

• Fix #210: Add scikit-learn plots to README. by @anevolbap in #307
• Add experiment design notation docs page to new Knowledge Base section of the docs by @drbenvincent in #312

Maintenance 🔧

• Upload versions due to Node 16 sunset. by @anevolbap in #310
• Fix #287 Update Makefile by @anevolbap in #309
• Drop support for Python 3.8 and 3.9, add 3.11 by @drbenvincent in #313

Full Changelog: 0.2.2...0.2.3

0.2.2

What's Changed

Bugfixes 🐛

• Fix #288: remove deprecated 'model' property in tests. by @anevolbap in #291
• Fix causal impact calculation error in synthetic control and interrupted time series by @drbenvincent in #296

Maintenance 🔧

• Expand data validation to also cover the scikit-learn experiment classes by @drbenvincent in #290

New Contributors

• @anevolbap made their first contribution in #291

Full Changelog: 0.2.1...0.2.2

0.2.1

What's Changed

New Features 🎉

• User specified number of significant figures for numbers in plots by @drbenvincent in #272
• Make results fully reproducible by @drbenvincent in #286
• User specified number of significant figures for numerical outputs by @drbenvincent in #283

Bugfixes 🐛

• changed splitting behaviour for PrePostFit by @rlaker in #275

Documentation 📖

• Fix missing images on PyPI page by @krz in #270

Maintenance 🔧

• Ignore causalpy/tests/ when running doctest to save a little time by @jpreszler in #280
• simplify doctest outputs by @jpreszler in #285

New Contributors

• @krz made their first contribution in #270
• @jsakv made their first contribution in #274
• @rlaker made their first contribution in #275

Full Changelog: 0.2.0...0.2.1

0.2.0

What's Changed

New Features 🎉

• Add ability to analyse Regression Kink Analysis Designs by @drbenvincent in #264

Full Changelog: 0.1.2...0.2.0