StatsPai Inc. @StatsPai

At this point, StatsPAI has already surpassed Stata and R in the causal inference domain and is beginning to reshape the Python ecosystem. The remaining step is to formalize key empirical results and publish them in a top statistical journal to provide strong academic validation.

StatsPAI 1.6.0: 250,000 lines of code, 925 functions I originally planned to release version 1.0.0, but ended up writing another 30,000 lines overnight and jumped straight to 1.6. This is simply the pace of the AI era—tools like this are being rapidly brought into existence.

keyboard_arrow_left Previous keyboard_arrow_right Next

StatsPAI is an agent-native causal inference and econometrics library for Python, with 900+ functions covering DiD, IV, RD, synthetic control, DML, Bayesian methods, causal discovery, and structural modeling. For both AI agents & human researchers. Aligned with Stata and R.

Repo: github.com/brycewang-stan…

🔗 Repo: github.com/brycewang-stan… PRs and forks very welcome — whether it's a new estimator, a bug fix, or a doc tweak. Star ⭐ if it saves you a regression. 🙏

StatsPAI v0.9.2 is out — biggest causal inference upgrade: • sp.decompose() unifies 18 methods (Oaxaca, RIF, FFL, Gelbach, Shapley…) • Synth: 13 estimators + sensitivity toolkit • v1.0 breadth pass: GRF, RD, spatial, causal discovery, survival, matching import statspai as sp

keyboard_arrow_left Previous keyboard_arrow_right Next

🔗 github.com/brycewang-stan… — forks, issues, PRs all welcome. ⭐ if it helps!

StatsPAI v0.9.1 — the RD release. sp.rd is now the most comprehensive RD toolkit in any language (Python/R/Stata): • 18+ estimators across 14 modules (~10.3k LOC) • Sharp · Fuzzy · Kink · 2D · Multi-cutoff · RDiT • CCT robust inference, CJM density tests, AK honest CIs!

🔗 Repo: github.com/brycewang-stan… PRs and forks very welcome — whether it's a new estimator, a bug fix, or a doc tweak. Star ⭐ if it saves you a regression. 🙏

🎯 StatsPAI v0.9.0 — the synthetic control upgrade. One module, now a full research stack: • 20 estimators (classic SC, SCM, ASC, MSC, RSC, PSC, SDID…) • 6 inference strategies (placebo, conformal, jackknife+) • one-click reports import statspai as sp → sp.synth.* 🐍

Meet CoPaper.AI — from Stanford's REAP program. Your AI research co-author. Faster. More rigorous. Covers everything from OLS, Logit, DiD, IV to causal forests. Full code. Reproducible papers. Built for economists & social scientists. 🔗 copaper.ai

23,000 AI Agent Skills for empirical research — open source. Causal inference, econometrics, literature review, paper reproduction — one command. Built by Stanford REAP. github.com/brycewang-stan…

keyboard_arrow_left Previous keyboard_arrow_right Next

When neural nets meet causal inference — get unbiased treatment effects from observational data. DeepIV in 3 lines: pip install StatsPAI import statspai as sp sp.deepiv(data=df, y="sales", treatment="price", instrument="fuel_cost") OLS lies. StatsPAI doesn't.

keyboard_arrow_left Previous keyboard_arrow_right Next

AI Tools for Economists Training Claude Code × Codex × CoPaper.AI × VS Code Data → Stata/Python → Paper, end-to-end 2 days / 12 hours / hands-on

keyboard_arrow_left Previous keyboard_arrow_right Next

Finally, a Stata Graph Editor clone for matplotlib. 29 themes, auto-generates Python code, one line to use. pip install statspai Open source. Your reproducible workflow just got a lot smoother.

keyboard_arrow_left Previous keyboard_arrow_right Next

Long-term vision: bring the mature causal inference & econometrics capabilities of Stata and R into a complete Python ecosystem. No more juggling statsmodels, linearmodels, EconML, and differences with incompatible APIs. One unified package for Python researchers.

Introducing StatsPAI — an open-source econometrics toolkit for the AI agent era. 150+ functions. One unified API. From OLS to causal forests. pip install statspai github.com/brycewang-stan…

keyboard_arrow_left Previous keyboard_arrow_right Next

What it covers: - Classic econometrics (IV, panel, Heckman) - DID, synthetic control, RDD - ML causal (DML, causal forest, TMLE) - Publication-ready output (.docx, .latex, .xlsx) Built for AI agents with standardized interfaces. MIT licensed, open source forever.

Introducing StatsPAI (Stanford REAP & CoPaper.AI): our vision is to build the unified Python toolkit for causal inference & econometrics. One import, every method: DID, synthetic control, IV, RDD, matching, and more. pip install statspai 💻 github.com/brycewang-stan…

keyboard_arrow_left Previous keyboard_arrow_right Next

🔬 We open-sourced 23,000+ agent skills for empirical research across 8 social science disciplines. Skills encode full research methodology — so AI agents know what a complete DID/IV/RDD analysis looks like, not just individual steps. github.com/brycewang-stan…