The Four-Step Process of Causal Inference
Welcome to Chapter 7!
This is a true milestone in our journey. In this chapter, we’ll learn how to neatly structure the entire causal inference process using the DoWhy library (Sharma & Kiciman, 2020). By the end of this chapter, you’ll be able to write production-ready causal inference pipelines using linear and non-linear estimators.
We’ll start with an introduction to DoWhy and its sister library, EconML (Battochi et al., 2019). After that, we’ll see how to use the graph modeling language (GML), which we introduced briefly in Chapter 4 to translate our assumptions regarding the data-generating process into graphs. Then, we’ll see how to compute causal estimands and causal estimates using DoWhy. Finally, we’ll introduce refutation tests and see how to apply them to our models. We’ll conclude the chapter with an example of a complete causal inference process. By the end of this chapter...
