DoWhy: An End-to-End Library for Causal Inference

Causal Data Science Meeting (https://causalscience.org/)

Download BibTex

In addition to efficient statistical estimators of a treatment’s effect, successful application of causal inference requires specifying assumptions about the mechanisms underlying observed data and testing whether they are valid, and to what extent. However, most libraries for causal inference focus only on the task of providing powerful statistical estimators. We describe DoWhy, an open-source Python library that is built with causal assumptions as its first-class citizens, based on the formal framework of causal graphs to specify and test causal assumptions. DoWhy presents an API for the four steps common to any causal analysis—1) modeling the data using a causal graph and structural assumptions, 2) identifying whether the desired effect is estimable under the causal model, 3) estimating the effect using statistical estimators, and finally 4) refuting the obtained estimate through robustness checks and sensitivity analyses. In particular, DoWhy implements a number of robustness checks including placebo tests, bootstrap tests, and tests for unoberved confounding. DoWhy is an extensible library that supports interoperability with other implementations, such as EconML and CausalML for the the estimation step.

GitHubGitHub

Related Tools

DoWhy: A library for causal inference

As computing systems are more frequently and more actively intervening in societally critical domains such as healthcare, education and governance, it is critical to correctly predict and understand the causal effects of these interventions. Without an A/B test, conventional machine learning methods, built on pattern recognition and correlational analyses, are insufficient for causal reasoning. Much like machine learning libraries have done for prediction, "DoWhy" is a Python library that aims to spark causal thinking and analysis. DoWhy provides a unified interface for causal inference methods and automatically tests many assumptions, thus making inference accessible to non-experts.

Access

Foundations of causal inference and its impacts on machine learning webinar

Many key data science tasks are about decision-making. They require understanding the causes of an event and how to take action to improve future outcomes. Machine learning (ML) models rely on correlational patterns to predict the answer to a question but often fail at these decision-making tasks, as the very decisions and actions they drive change the patterns they rely on. Causal inference methods, in contrast, are designed to rely on patterns generated by stable and robust causal mechanisms, even as decisions and actions change. With insights gained from causal methods, the new, growing field of causal machine learning promises to address fundamental ML challenges in generalizability, interpretability, bias, and privacy.

In this webinar, join Microsoft researchers Amit Sharma and Emre Kıcıman to learn about the fundamentals of causal inference. You will learn how a target question of cause and effect can be captured in a formal graphical model and answered systematically using available data. The researchers will introduce a four-step causal modeling framework for analyzing decision-making tasks and walk-through code examples using the DoWhy Python library that implements the framework. You will also discover how causal methods can be useful to improve ML models in terms of their generalizability, explainability, fairness, and robustness.

Together, you’ll explore:

  • Why causal reasoning is necessary for decision-making
  • The difference between a prediction and a decision-making task
  • How the DoWhy library can help you conduct a robust causal inference analysis by translating domain knowledge to a causal graph and validating the graph using available data
  • The connections between causal inference and the challenges of modern ML models

Resource list:

*This on-demand webinar features a previously recorded Q&A session and open captioning.

Explore more Microsoft Research webinars: https://aka.ms/msrwebinars (opens in new tab)