Causal Inference in Python

Chapter 12. Next Steps

It has been a long way since you were first introduced to counterfactuals. This book has taken you on a journey through the world of causal inference, starting with the basics and gradually building up to more advanced concepts and techniques. You should now have a solid understanding of how to reason about causation and how to use various methods to untangle causation from correlation in your data.

You have learned about the importance of A/B testing as the gold standard for causal inference, the power of graphical models for causal identification, and the use of linear regression and propensity weighting for bias removal. You have explored the intersection between machine learning and causal inference and how to use these tools for personalized decision making.

Furthermore, you have learned how to incorporate the time dimension into your causal inference analyses using panel datasets and methods like difference-in-differences and synthetic control. Finally, you have gained an understanding of alternative experiment designs for when randomization is not possible, such as geo and switchback experiments, instrumental variables, and discontinuities.

With the knowledge and tools presented in this book, you are equipped to tackle real-world problems and make informed decisions based on causation rather than correlation. I hope you enjoyed it and that it keeps being useful to you throughout your career.

This being an introductory book, I intentionally left out ...