alterlab-statsmodels — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited alterlab-statsmodels (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Statsmodels is Python's premier library for statistical modeling, providing tools for estimation, inference, and diagnostics across a wide range of statistical methods. Apply this skill for rigorous statistical analysis, from simple linear regression to complex time series models and econometric analyses.
This skill should be used when:
| Scenario | Use Instead |
|---|---|
| Choosing which test fits the design, checking assumptions, and writing APA-style results | alterlab-statistical-analysis |
| Bayesian or hierarchical models with posterior distributions (PyMC, NUTS) | alterlab-pymc |
| Quasi-experimental causal designs (DiD, IV, RD, panel fixed effects, event studies) as a full workflow | alterlab-causal-inference |
| Predictive machine learning with cross-validated tuning rather than coefficient inference | alterlab-scikit-learn |
| Zero-shot forecasting with a pretrained foundation model instead of fitting ARIMA/ETS | alterlab-timesfm |
statsmodels 0.15.0 (Aug 2026) is the first release since 0.14 (2023). Changes that affect everyday code:
adfuller, kpss, acf (with qstat/alpha), het_arch, acorr_lm, acorr_breusch_godfrey, het_goldfeldquandt and a few others still return the legacy tuple but emit a FutureWarning; pass result_object=True to get the named result now (e.g. adfuller(y, result_object=True).pvalue, het_arch(r, result_object=True).lmpval). The default switches in 0.16.seed= / random_state= arguments are deprecated in favor of rng= (SPEC 7).grangercausalitytests(verbose=...), AutoReg(old_names=...); kpss(nlags=None) now raises — pass 'auto', 'legacy', or an integer.SM_FORMULA_ENGINE=formulaic. Model and formula APIs also accept Polars DataFrames.pairwise_tukeyhsd(..., use_var="unequal")), the Leybourne-McCabe stationarity test (statsmodels.tsa.stattools.leybourne), MultivariateLS, and robust MM/S estimators.Copy-paste starting points for OLS, logistic regression, ARIMA, GLM/Poisson, and the R-style formula API are in references/quickstart_examples.md. Core rule: always sm.add_constant() for an intercept unless you deliberately want none.
Comprehensive suite of linear models for continuous outcomes with various error structures.
Available models:
Key features:
When to use: Continuous outcome variable, want inference on coefficients, need diagnostics
Reference: See references/linear_models.md for detailed guidance on model selection, diagnostics, and best practices.
Flexible framework extending linear models to non-normal distributions.
Distribution families:
Link functions:
Key features:
When to use: Non-normal outcomes, need flexible variance and link specifications
Reference: See references/glm.md for family selection, link functions, interpretation, and diagnostics.
Models for categorical and count outcomes.
Binary models:
Multinomial models:
Count models:
Key features:
When to use: Binary, categorical, or count outcomes
Reference: See references/discrete_choice.md for model selection, interpretation, and evaluation.
Comprehensive time series modeling and forecasting capabilities.
Univariate models:
Multivariate models:
Advanced models:
Key features:
When to use: Time-ordered data, forecasting, understanding temporal dynamics
Reference: See references/time_series.md for model selection, diagnostics, and forecasting methods.
Extensive testing and diagnostic capabilities for model validation.
Residual diagnostics:
Influence and outliers:
Hypothesis testing:
Multiple comparisons:
Effect sizes and power:
Robust inference:
When to use: Validating assumptions, detecting problems, ensuring robust inference
Reference: See references/stats_diagnostics.md for comprehensive testing and diagnostic procedures.
smf.ols, smf.logit, smf.poisson, interactions, C(), I()) → references/quickstart_examples.md.references/model_selection.md.references/workflows_and_practices.md.references/linear_models.md.references/glm.md.references/discrete_choice.md.references/time_series.md.references/stats_diagnostics.md.references/model_selection.md.references/quickstart_examples.md — copy-paste OLS / Logit / ARIMA / GLM examples and the R-style formula API.references/linear_models.md — OLS, WLS, GLS, GLSAR, quantile, mixed effects, recursive/rolling; diagnostics, influence, robust SEs, hypothesis testing.references/glm.md — all distribution families, link functions, interpretation, pseudo R-squared, residual analysis.references/discrete_choice.md — binary (Logit/Probit), multinomial, count (Poisson/NB/ZIP/ZINB/hurdle), ordinal, marginal effects.references/time_series.md — AR/ARIMA/SARIMAX/ETS, VAR/VARMAX/dynamic factor, state space, stationarity, forecasting, Granger/IRF/FEVD.references/stats_diagnostics.md — residual diagnostics, influence/outliers, parametric and non-parametric tests, ANOVA, multiple comparisons, robust covariances, power/effect sizes.references/model_selection.md — AIC/BIC comparison, likelihood ratio test, cross-validation.references/workflows_and_practices.md — best practices, end-to-end workflows, common pitfalls, search patterns, official docs links.Part of the AlterLab Academic Skills suite.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.