Bayesian statistical modeling with PyMC v5+. Use when building probabilistic models, specifying priors, running MCMC inference, diagnosing convergence, or comparing models. Covers PyMC, ArviZ, pymc-bart, pymc-extras, nutpie, and JAX/NumPyro backends. Triggers on tasks involving: Bayesian inference, posterior sampling, hierarchical/multilevel models, GLMs, time series, Gaussian processes, BART, mixture models, prior/posterior predictive checks, MCMC diagnostics, LOO-CV, WAIC, model comparison, or causal inference with do/observe.
A Claude Code skill that provides comprehensive guidance for Bayesian statistical modeling with PyMC v5+. When loaded, Claude gains expert knowledge of PyMC workflows, best practices, and common patterns.
When you work on PyMC modeling tasks, this skill is automatically loaded to provide:
Copy or symlink this directory to your Claude Code skills folder:
# Create skills directory if it doesn't exist
mkdir -p ~/.claude/skills
# Clone or copy this repository
git clone https://github.com/fonnesbeck/pymc-modeling.git ~/.claude/skills/pymc-modeling
The skill will automatically activate when Claude detects tasks involving Bayesian inference, PyMC, ArviZ, or related topics.
SKILL.md # Main skill: model specification, inference, diagnostics
CLAUDE.md # Instructions for Claude when editing this skill
references/
arviz.md # Comprehensive ArviZ guide (all plots, interpretation)
priors.md # Prior selection (weakly informative defaults)
inference.md # MCMC backends (nutpie, NumPyro/JAX, approximate)
diagnostics.md # Quick diagnostics reference, troubleshooting
gotchas.md # Common pitfalls and performance issues
gp.md # Gaussian processes (HSGP, kernels, priors)
timeseries.md # Time series (AR, structural, seasonality)
bart.md # BART (pymc-bart) usage and interpretation
mixtures.md # Mixture models and label switching
specialized_likelihoods.md # Zero-inflated, censored, ordinal, robust
custom_models.md # Custom distributions (DensityDist, Potential)
pm.DensityDist and pm.CustomDistpm.Potentialpm.Simulatorpm.do and pm.observeWhen this skill is loaded, Claude will follow these principles:
| Topic | Recommendation | |-------|----------------| | Sampler | nutpie by default (2-5x faster), PyMC NUTS as fallback | | Parameterization | Non-centered for hierarchical models with weak data | | GPs | HSGP for n > 500 points; full GP only for small datasets | | Diagnostics | Check r_hat < 1.01, ESS > 400, zero divergences before interpretation | | Notebooks | marimo preferred over Jupyter |
The skill assumes these packages are available:
pymc >= 5.0arviznutpie (recommended)pymc-bart (for BART models)pymc-extras (for specialized distributions)Optional for GPU/JAX acceleration:
numpyrojaxThis skill is provided as-is for use with Claude Code.
Search for places (restaurants, cafes, etc.) via Google Places API proxy on localhost.
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Start voice calls via the OpenClaw voice-call plugin.
Notion API for creating and managing pages, databases, and blocks.
Gemini CLI for one-shot Q&A, summaries, and generation.
Category:developer