v0.6.0: Complete Econometrics Toolkit + JOSS Paper

StatsPAI v0.6.0

New Methods (30+ modules)

• GLM & Discrete Choice: glm(), logit(), probit(), mlogit(), ologit(), oprobit(), clogit()
• Count Data: poisson(), nbreg(), ppmlhdfe(), zip_model(), zinb(), hurdle()
• Advanced IV: liml(), jive(), lasso_iv()
• Survival Analysis: cox(), kaplan_meier(), survreg(), logrank_test()
• Time Series: var(), granger_causality(), irf(), structural_break(), engle_granger(), johansen()
• Nonparametric: lpoly(), kdensity()
• Experimental Design: randomize(), balance_check(), attrition_test(), optimal_design()
• Missing Data: mice(), mi_estimate()
• Mendelian Randomization: mendelian_randomization(), mr_egger(), mr_ivw()
• Panel Extensions: panel_logit(), panel_probit(), panel_fgls(), interactive_fe(), panel_unitroot()
• System Estimation: sureg(), three_sls(), gmm()
• Structural: blp(), frontier()
• Other: fracreg(), betareg(), biprobit(), etregress(), truncreg(), mixed()

Smart Workflow Engine (unique to StatsPAI)

• recommend() — suggests estimators given data + research question
• compare_estimators() — runs multiple methods, reports agreement
• assumption_audit() — tests all assumptions in one call
• sensitivity_dashboard() — multi-dimensional sensitivity analysis
• pub_ready() — journal-specific publication readiness checklist
• replicate() — built-in famous datasets (Card 1995, LaLonde 1986, Lee 2008)

JOSS Paper

• Paper submitted to Journal of Open Source Software
• Validated against published benchmarks (Card 1995, Dehejia & Wahba 1999, Lee 2008)
• Cross-validated against EconML

Total: 390+ functions from a single import statspai as sp