Start free trial
← Causal Inference 101

5 min read

Causal discovery basics

When you do not know the causal graph upfront — algorithms that suggest structure from data, and how to use them carefully.

Causal discovery algorithms (PC, LiNGAM, and others) search for statistical patterns consistent with causal structures. They help brainstorm hypotheses when domain knowledge is thin.

Discovery output is not proof. Confounders, measurement error, and small samples can mislead. Treat discovered edges as proposals to validate with experts and follow-up experiments.

Best practice

  • Combine data-driven graphs with subject-matter knowledge.
  • Use discovery to suggest confounders and mediators, not as a black box oracle.
  • Re-run effect estimation with explicit methods after narrowing the graph.

Run this method on your data — no Python

CausalLens runs matching, DiD, causal forests, DoWhy refutation, and more — with balance tables, sensitivity checks, and PDF export.

Start 14-day Pro trialAll methodsPricing