Measuring Causal Relations with Python

In the realm of data science and statistical analysis, the study of causal inference and causal discovery has gained tremendous importance. Understanding the causal relationships between variables is essential for drawing accurate conclusions from data, predictive modeling, and decision-making in a wide array of domains.

In this project, we’ll explore causal inference and causal discovery, employing Python and libraries like NetworkX, gCastle, and DoWhy. We’ll explore various methods and techniques for identifying and measuring causal relations between variables. We’ll start with the fundamentals—creating causal graphs and identifying confounders and colliders. Next, we’ll explore traditional techniques like linear regression and instrumental variables, building upon the knowledge extracted from the graphs. We’ll then be ready to move on to more advanced techniques like calculating propensity scores and using doubly robust estimators and causal forests.