statsmodels

Fit statistical models and produce inference and diagnostics for tabular datasets.

2|Updated Apr 12, 2026
One-click install
npx skills add https://github.com/Clay-HHK/claude-config --skill statsmodels-clay-hhk
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: statsmodels
Source: https://github.com/Clay-HHK/claude-config/tree/main/skills/statsmodels
Command: npx skills add https://github.com/Clay-HHK/claude-config --skill statsmodels-clay-hhk

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Provide rigorous estimation, inference, and diagnostics for statistical models so analysts can make valid conclusions, compare specifications, and produce reproducible forecasts without ad-hoc scripting.

Core Features & Use Cases

  • Modeling variety: OLS, WLS, GLS, quantile regression, mixed effects, GLMs (logit, Poisson, Gamma, Tweedie), discrete choice (logit, probit, MNLogit), count models (Poisson, Negative Binomial, ZIP/ZINB), and time series (ARIMA, SARIMAX, VAR, state space).
  • Inference & diagnostics: Robust and cluster standard errors, marginal effects, hypothesis tests, residual and influence diagnostics, model comparison via AIC/BIC and likelihood-ratio tests, and forecast intervals.
  • Practical workflows: Econometrics and policy inference, clinical/biostatistics modeling, count-data analysis, and time series forecasting with publication-ready tables and plots using the formula API.

Quick Start

Fit an OLS model with robust standard errors and produce a formatted coefficient table, p-values, confidence intervals, and residual diagnostics for the dataframe df with outcome y and predictors x1 and x2.

Frequently Asked Questions about statsmodels

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run OLS regression with robust standard errors and get diagnostic outputs?

Fit OLS regression with robust standard errors by applying the formula API to your tabular dataset, yielding coefficient estimates, p-values, confidence intervals, and residual diagnostics.

Can I model count data using Poisson or Negative Binomial regression?

Yes, you can model count data using Poisson, Negative Binomial, and zero-inflated models, producing coefficient estimates, marginal effects, and AIC/BIC comparisons for valid inference.

Does this support time series forecasting with ARIMA and SARIMAX models?

Time series forecasting is supported through ARIMA, SARIMAX, VAR, and state space models, generating forecasts with prediction intervals for econometric and biostatistical workflows.

What's the best way to compare statistical model specifications for econometrics?

Compare statistical model specifications using AIC/BIC metrics and likelihood-ratio tests, evaluating residual outputs and influence diagnostics to ensure valid econometric conclusions.

Can I estimate generalized linear models like logit and probit for discrete choice?

Estimate generalized linear models including logit, probit, and MNLogit for discrete choice analysis, returning coefficient estimates, marginal effects, and clustered standard errors.

When should I use quantile regression instead of standard OLS regression?

Use quantile regression instead of OLS when modeling conditional quantiles is necessary, producing robust coefficient estimates and inference outputs for tabular datasets.