Writing

A conceptual introduction to causal graphs.

Why do we need causal inference? What's wrong with just plotting X vs. Y and reading the answer?

A brief introduction to Pearlian causality and the back-door criterion

What are the tradeoffs when you do causal instead of correlative analysis? Is causal inference practical on a deadline?

We can use the tools we've developed to look at our standard approaches in data science, and realize sometimes we're not measuring what we think we're measuring!

How can we do observational causal inference in the usual data science workflow?