A "Quick", Broad Introduction to Causal Inference

(from an elementary, mostly observational point-of-view)

For a clean view from the NBViewer, click here

This IPython notebook was used to give the Penn Medicine Predictive Healthcare team exposure to the topic of causal inference and introduce them to some established methods. This notebook can be presented as a Reveal.js slideshow using the RISE extension. Feel free to use it under the terms of the Creative Commons Attribution 4.0 International License. Any suggestions for improvement are most welcome.

Outline

• The What and Why of Causality
• The Need for Causal Inference on Observational Data
• Qualitative Method
• Hill Criterion
• Some Established Formal Methods
• Matched Sampling (Rubin et al.)
• Graphical Models (Pearl, Spirtes et al.)
• Granger Causality (Granger)
• Possible Shortcomings
• Books on Causal Inference
• Citations

Special thanks to Pedro Ortega for allowing me to use some of his slides.

This work is licensed under a Creative Commons Attribution 4.0 International License.