A Gentle Introduction To Causal Inference
summary
Claims that causal relationships are recoverable from observational data without running RCTs, demonstrates with a 10,000‑patient example where the population effect is −14% while both gender cohorts show +10% each, explains resolving that contradiction with a weighted average to get the ACE=average causal effect, introduces graph models to justify which variables to control for, and shows working artifacts — Jupyter notebooks and a Streamlit calculator — so the methods and numbers are reproducible and inspectable.
AI-written summary · about
chapters (9)
- 0:00Causal inference introduction
- 3:13Correlations vs observations
- 8:35Simpson's Paradox examples
- 18:26Resolving Simpson's Paradox
- 20:31Graph models for causality
- 23:19Causal inference assumptions
- 24:21Pearl's ladder and counterfactuals
- 26:51Summary and resources
- 28:36Q&A on graph models
related talks
41mOn the Path to Causal Inference
24mA Friendly Introduction To Causality
24mAn introduction to Causal Inference with Python – making accurate estimates of cause and effect from
01h 01mThe Mathematics of Causal Inference, with Reflections on Machine Learning and the Logic of Science
45mCausality and Statistics, a Love Not So Impossible After All
31mPython and the Holy Grail of Causal Inference
29m🧠🧹 Causality - Mental Hygiene for Data Science
39mCausal Effect Estimation in Practice: Lessons Learned from E-commerce & Banking