How does Statsmodels differ from scikit-learn?

While scikit-learn focuses primarily on predictive accuracy, Statsmodels emphasizes statistical inference, providing detailed p-values, confidence intervals, and diagnostic tests to explain relationships between variables.

Does this skill support R-style formulas?

Yes, it fully supports the Formula API via Patsy, allowing you to define models using the intuitive 'y ~ x1 + x2' syntax directly with Pandas DataFrames.

Can I use this for time series forecasting?

Absolutely. It includes robust tools for stationarity testing (ADF), ACF/PACF plots, and complex modeling using ARIMA, SARIMAX, and VAR models.

What model diagnostics are included?

It provides automated checks for heteroscedasticity (Breusch-Pagan), normality of residuals (Omnibus/Jarque-Bera), multicollinearity (VIF), and outlier detection via Cook's Distance.

How does it handle categorical data?

By using the C() operator in formulas, the skill automatically handles dummy variable encoding and ensures the reference category is handled correctly to avoid the dummy variable trap.

Statsmodels Statistical Analysis

Name: Statsmodels Statistical Analysis
Author: tondevrel

bytondevrel

•

データサイエンスとML

Performs advanced statistical modeling, hypothesis testing, and rigorous data inference using R-style formulas.

This skill empowers Claude to conduct professional-grade statistical analysis by bridging Python with the mathematical rigor of R. It provides specialized capabilities for estimating complex models like OLS, GLM, and SARIMAX, performing exhaustive diagnostic tests such as normality and heteroscedasticity checks, and generating detailed summary tables required for academic and scientific reporting. Whether you are analyzing time series data, conducting survival analysis, or testing experimental hypotheses, this skill ensures statistical validity through automated residual analysis, multicollinearity detection, and proper model specification.

主な機能

01Advanced time series analysis including ARIMA, SARIMAX, and state space models

021 GitHub stars

03R-style formula support for intuitive model specification using Patsy

04Generalized Linear Models (GLM) for logistic, Poisson, and Gamma regression

05Comprehensive diagnostic tools for residual analysis and hypothesis testing

06Detailed statistical summary tables with p-values, confidence intervals, and AIC/BIC

ユースケース

01Conducting academic-grade hypothesis testing and ANOVA for experimental data

02Building econometrics models for market attribution and causal inference

03Forecasting complex time series data with seasonal and exogenous variables

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add tondevrel/scientific-agent-skills statsmodels

For use in Claude.ai and ChatGPT

Download Skill