Explain causal inference and its applications in data science.

Causal inference determines cause-and-effect relationships beyond correlation, crucial for decision-making and policy evaluation.\n\n• Methods: Randomized experiments, instrumental variables, regression discontinuity\n• Frameworks: Potential outcomes, directed acyclic graphs (DAGs)\n• Challenges: Confounding variables, selection bias, external validity\n• Applications: Marketing attribution, policy evaluation, treatment effects\n\nExample: Measuring marketing campaign effectiveness uses difference-in-differences design, controls for seasonal trends and external factors, and estimates causal impact on sales using synthetic control methods and robustness checks.