Deconfounding like a pro with DoWhy

In an earlier article, I made the case for causality, set a bit of theoretical ground, and argued that a solid understanding of causal mechanisms should be in the toolset of every data scientist informing key decisions through data. A causal inference analysis often entails drawing a graph of what may be causing what, identifying confounders, and stratifying those to find the effect of a treatment on an outcome. Doing this properly allows you to stay clear of spurious correlations and absurd claims. Finding the right confounders, which are the factors that influence both a treatment and an outcome is therefore key in solving for the causal effects. DoWhy is one of the most powerful libraries for the task, completely open-source by the Microsoft research team.