stan-dev/bayesplot

bayesplot R package for plotting Bayesian models

What it solves

Bayesian modeling often produces complex posterior draws and MCMC (Markov Chain Monte Carlo) data that are difficult to interpret visually. bayesplot provides a standardized library of plotting functions to help users diagnose model performance and visualize the results of Bayesian models.

How it works

The package is built as an R library that integrates with ggplot2. It takes posterior draws (typically from MCMC) and transforms them into a variety of specialized plots. It supports several types of visualizations:

  • Posterior Draws: Visualizing the distribution of parameters.
  • MCMC Diagnostics: Identifying issues with chain convergence and sampling efficiency.
  • HMC/NUTS Diagnostics: Specific tools for diagnosing Hamiltonian Monte Carlo (HMC) and No-U-Turn Sampler (NUTS) performance, such as energy plots and divergence scatter plots.
  • Predictive Checking: Graphical posterior or prior predictive checks to compare model predictions against observed data.

Who it’s for

It is designed for researchers and developers working with Bayesian modeling, particularly those using the Stan ecosystem (such as rstan or rstanarm), though it can be used with any Bayesian model that produces posterior draws.

Highlights

  • ggplot2 Integration: All plots are ggplot objects, allowing for full customization using standard ggplot2 functions.
  • Comprehensive Diagnostics: Includes specific tools for MCMC and NUTS energy diagnostics.
  • Predictive Analysis: Built-in functions for posterior predictive checks (PPC).
  • Developer-Friendly: Provides a common set of functions that other Bayesian modeling packages can use as a backend for visualization.

Related

  • Project
    easystats/bayestestR

    An R package that provides a comprehensive set of functions to analyze and describe posterior distributions from Bayesian models, focusing on centrality, uncertainty, and effect significance.

  • Project
    easystats/performance

    An R package for assessing, comparing, and testing the quality of statistical regression models through fit indices and diagnostic checks.

  • Project
    paul-buerkner/brms

    brms is an R package that provides a high‑level, lme4‑style interface for fitting Bayesian generalized (non‑linear) multilevel models using Stan. It supports many response families, flexible priors, and a suite of post‑fit diagnostics and visualizations, enabling users to perform sophisticated Bayesian regression without writing Stan code directly.

  • Project
    geomstats/geomstats

    A Python package for computations, statistics, and machine learning on manifolds, enabling the analysis of non-Euclidean data like rotations and shapes.

  • Project
    erdogant/bnlearn

    bnlearn is a Python library for learning, parameterising, and querying Bayesian networks. It supports structure discovery (various scores and search methods), CPT estimation, causal inference (do‑calculus), synthetic data generation, and visualisation, all wrapped in a simple API.