pymc-labs/CausalPy

A Python package for causal inference in quasi-experimental settings

What it solves

CausalPy provides a framework for estimating causal effects in quasi-experimental designs where randomized controlled trials are not possible. It addresses the need for research-grade, reproducible causal inference workflows that move beyond simple point estimates to include full uncertainty quantification.

How it works

The library implements a variety of quasi-experimental methods (such as Difference-in-Differences, Synthetic Control, and Regression Discontinuity) using either Bayesian estimation via PyMC or traditional Ordinary Least Squares (OLS) via scikit-learn. It focuses on a "diagnostics-first" approach, providing tools to calculate Highest Density Intervals (HDI) for Bayesian models and the Region of Practical Equivalence (ROPE) to determine if an effect is practically significant.

Who it’s for

It is designed for researchers and data scientists who have a plausible quasi-experimental design (e.g., policy changes, threshold rules, or staggered rollouts) and require rigorous, uncertainty-aware estimates for high-stakes decision-making.

Highlights

  • Diverse Method Palette: Supports Synthetic Control, Geographical Lift, ANCOVA, Difference-in-Differences (including staggered), Interrupted Time Series, Regression Discontinuity/Kink, Instrumental Variables, and Inverse Propensity Score Weighting.
  • Bayesian-First Approach: Leverages PyMC for full uncertainty quantification and posterior estimates.
  • Decision-Ready Outputs: Generates effect summaries with credible intervals and publication-quality plots.
  • Practical Significance Testing: Includes ROPE analysis to assess if effects exceed meaningful thresholds.

Related

  • Project
    pgmpy/pgmpy

    A Python library for causal and probabilistic reasoning using graphical models, providing tools for structure learning, parameter estimation, and causal inference.

  • Project
    igerber/diff-diff

    A Python library for Difference-in-Differences causal inference that provides a wide range of econometric estimators and diagnostic tools for measuring policy and campaign lift.

  • Project
    py-why/dowhy

    DoWhy is a Python library for end‑to‑end causal inference. It lets you define a causal graph, automatically identifies the causal effect, estimates it with a variety of statistical methods, and runs refutation tests to check robustness. The library also supports graphical causal models for root‑cause analysis, counterfactuals, and interventional sampling. Install via `pip install dowhy` and follow the four‑step workflow (model → identify → estimate → refute).

  • Project
    uber/causalml

    A Python package for uplift modeling and causal inference that estimates the causal impact of interventions on individual users to optimize targeting and personalization.

  • Project
    erdogant/bnlearn

    bnlearn is a Python library for learning, parameterising, and querying Bayesian networks. It supports structure discovery (various scores and search methods), CPT estimation, causal inference (do‑calculus), synthetic data generation, and visualisation, all wrapped in a simple API.