StatsPAI_skill

What problem does it solve?

It helps you estimate treatment effects and run robust causal inference without stitching together many separate tools, by guiding an end-to-end empirical workflow in Python from diagnostics to estimation to robustness.

Core Features & Use Cases

• Full empirical causal pipeline: descriptive stats and balance checks, pre-flight diagnostics, estimand-first research question specification, estimator selection, estimation, and robustness.
• Estimand-first decisioning: uses a research-question DSL to formalize population, treatment, outcome, estimand, and design choices (e.g., “DID vs RD vs IV?”).
• LLM-assisted DAG discovery: proposes, validates, and constrains causal DAGs to inform identification reasoning.
• Broad method coverage: OLS, IV, DID (including staggered-DID workflows), RDD, PSM, SCM, modern ML causal inference (DML, causal forests, meta-learners, TMLE), and text-as-treatment.
• Structured, agent-friendly outputs: returns self-describing result objects with summaries, diagnostics, and export/citation helpers.

Quick Start

Ask the AI to: run a full StatsPAI causal analysis for my dataset to estimate the treatment effect of training on wage using DID, including balance/pre-flight diagnostics, the estimand-first plan, estimation, and robustness checks.