Causal Inference in Python by Matheus Facure

Prediction tells you what will happen. Causal inference tells you what happens if you intervene — and that is the question every pricing, discount and product decision actually asks. Matheus Facure’s book teaches the methods that answer it, in Python, on the kinds of problems a technology company genuinely faces.

What the book covers

• Association versus causation, potential outcomes and the causal graph vocabulary
• Randomised experiments, and what to do when you cannot run one
• Regression and propensity score methods for observational data
• Difference-in-differences and synthetic control for geographic and time-based rollouts
• Panel data methods and two-way fixed effects
• Heterogeneous treatment effects and causal machine learning for personalisation
• Instrumental variables, regression discontinuity and non-compliance

Who it’s for

Data scientists and analysts in industry, economists moving into tech, and ML engineers whose models are being used to make decisions rather than forecasts. Comfortable Python and basic statistics assumed.

Edition details

• Author: Matheus Facure
• Publisher: O’Reilly Media
• Edition: 1st Edition
• Format: Paperback
• ISBN: 978-1098140250
• Language: English

Ships free within the United States from Shelfcove.