Example notebooks

These examples are also available on GitHub. You can run them locally after cloning DoWhy and installing Jupyter. Or you can run them directly in a web browser using the Binder environment.

Getting started

• Getting started with DoWhy: A simple example
• Confounding Example: Finding causal effects from observed data
• DoWhy: Different estimation methods for causal inference
• Testing Assumptions in model with DoWhy: A simple example
• Simple example on using Instrumental Variables method for estimation
• Different ways to load an input graph
• DoWhy: Interpreters for Causal Estimators
• Demo for the DoWhy causal API
• Do-sampler Introduction
• Basic Example for GCM-Based Inference

Using benchmark datasets

• DoWhy example on ihdp (Infant Health and Development Program) dataset
• DoWhy example on the Lalonde dataset
• Applying refutation tests to the Lalonde and IHDP datasets
• Lalonde Pandas API Example

Advanced

• Conditional Average Treatment Effects (CATE) with DoWhy and EconML
• Mediation analysis with DoWhy: Direct and Indirect Effects
• A Simple Example on Creating a Custom Refutation Using User-Defined Outcome Functions
• Estimating effect of multiple treatments
• Iterating over multiple refutation tests
• Identifying Effect using ID Algorithm
• Root cause analysis (RCA) of latencies in a microservice architecture
• Root causes of changes in a supply chain
• Impact of 401(k) eligibility on net financial assets