PyWhy

DoWhy 0.14

Estimate the average effect of a binary treatment

Open in Studio

What to provide

Provide 20 to 256 rows with three values each. Include a numeric pre-treatment confounder, treatment as 0 or 1, and a numeric outcome.

How it works

DoWhy identifies an effect under the declared causal graph. Studio fits a linear regression and returns the estimated average effect and the count for each group.

How to read the result

Use it when both groups have comparable observations and the confounder represents relevant common causes. Causal interpretation depends on the graph, adjustment and available data.

Credits and licenses

DoWhy 0.14 · PyWhyMIT
Official documentation