Getting Started

What CausalML is, an introduction to causal machine learning for readers new to it, how to install the package, a tour of the API in runnable snippets, an end-to-end walkthrough of estimating and validating heterogeneous treatment effects on a benchmark, and answers to common questions.

• About CausalML
  • GitHub Repo
  • Mission
  • Contributing and Governance
• Intro to Causal Machine Learning
  • What is Causal Machine Learning?
  • Difference from Traditional Machine Learning
  • Measuring Causal Effects
  • Why a Predictive Model Is Not Enough
  • Example Use Cases
• Installation
  • System Requirements
  • Install using conda
  • Install from PyPI
  • Install using uv
  • Install from source
  • Windows
  • Running Tests
• API Quickstart
  • Propensity Score
  • Average Treatment Effect (ATE) Estimation
  • More algorithms
  • Interpretation
  • Validation
  • Synthetic Data Generation Process
  • Sensitivity Analysis
  • Feature Selection
• Estimating and Validating Heterogeneous Treatment Effects
  • Step 1: State the question
  • Step 2: Load the data
  • Step 3: State the identification and check overlap
  • Step 4: Train four estimators
  • Step 5: Evaluate against the ground truth
  • Step 6: Validate as if the truth were unknown
  • Step 7: Visualize the ranking with gain and TOC curves
  • Step 8: Stress the assumptions
  • Step 9: The replication protocol
  • Summary
  • Where to next
• Frequently Asked Questions
  • Importing CausalML fails with an XGBoost error on macOS
  • Downloading a benchmark dataset fails with CERTIFICATE_VERIFY_FAILED
  • My causal tree results changed after upgrading
  • UpliftRandomForestClassifier got slower after upgrading
  • My Qini or AUUC score is negative
  • My propensity scores pile up near 0 or 1
  • The Twins dataset’s outcome looks like earnings, not mortality
  • My IHDP results differ across replications more than expected
  • Do I need TensorFlow, PyTorch or JAX?