/CausalInference

Primary LanguageJupyter NotebookMIT LicenseMIT

GitHub Twitter @data4sci GitHub top language GitHub repo size GitHub last commit Binder

Graphs For Science Visualization For Science Sunday Briefing

Why and What If: Causal Inference for Everyone

Code and slides to accompany the online series of webinars: https://data4sci.com/causality by Data For Science.

Run the code in Binder: Binder

How do causes lead to effects? Can you associate the cause leading to the observed effect? Big Data opens the doors for us to be able to answer questions such as this, but before we are able to do so, we must go beyond classical probability theory and dive into the field of Causal Inference.

In this course, we will explore the three steps in the ladder of causation: 1. Association 2. Intervention 3. Counterfactuals with simple rules and techniques to move up the ladder from simple correlational studies to fully causal analyses. We will cover the fundamentals of this powerful set of techniques allowing us to answer practical causal questions such as “Does A cause B?” and “If I change A how does that impact B?”

Schedule

1. Approaches to Causality

  • Probability Theory
  • Simpsons Paradox
  • A/B Testing
  • Granger Causality
  • Graphical Models
  • The Ladder of Causality

2. Properties of Graphical models

  • Chains
  • Forks
  • Colliders
  • d-separation

3. Interventions

  • Back-door criterion
  • Front-door criterion
  • Mediation

4. Counterfactuals

  • The fundamental laws of counterfactuals
  • Graphical representation
  • Practical Applications
  • Connections to Machine Learning

Slides: http://data4sci.com/landing/causality/