---
id: arviz-stats
version: "1.3.0"
license: unclear
license_treatment: permissive
maintenance: active
---

# arviz-stats — Statistical computation and diagnostics for ArviZ.

License: permissive · Maintenance: active · Downloads: 730.9K/mo

## What it is and what it does

arviz-stats is the statistical computation subpackage within the ArviZ ecosystem for Bayesian model analysis. It handles the core diagnostics and statistical summaries needed after running Bayesian inference, such as convergence checks, posterior summaries, and model comparison metrics. The package is designed as a modular component that can be used standalone with just numpy and scipy, or integrated into the broader ArviZ suite with xarray support for richer data structures.

The package targets both researchers doing exploratory Bayesian analysis and library developers who want to compute diagnostics without adopting the full ArviZ dependency tree. It supports current Python versions (3.12+) and maintains active development with regular updates.

Use it for:

- Computing convergence diagnostics (Rhat, effective sample size) for MCMC chains after inference
- Generating posterior predictive checks and model comparison statistics for Bayesian model validation
- Integrating Bayesian diagnostics into custom inference pipelines without adopting xarray
- Analyzing posterior samples from Stan, PyMC, or other Bayesian inference libraries
- Building downstream tools that need standardized Bayesian statistical summaries

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

Provides statistical computations and diagnostic functions for Bayesian model analysis, including posterior analysis, model checking, and comparison metrics.

Yes, if you work with Bayesian models and need standard diagnostics and posterior analysis. The low dependency footprint (numpy, scipy only) and active maintenance make it a practical choice. Install with the optional xarray extra for full feature access, or use the minimal version if you're integrating into a library with strict dependencies.

## Install

pip install arviz-stats
uv add arviz-stats
poetry add arviz-stats

## Installing arviz-stats

Before you install: Low friction installation with only numpy and scipy as runtime dependencies. Active maintenance with a release 3 days old and recent commits.

License in practice: Licensed under Apache License (permissive), allowing broad use in commercial and private projects with minimal restrictions.

Quickstart:

pip install arviz-stats

import arviz_stats
# Use statistical functions for Bayesian model diagnostics
# Functions accept numpy arrays or scipy distributions

Requires Python 3.12 or later. Optional xarray dependency unlocks full feature set; minimal install provides only array-based functions.

Verify before relying:

- Specific diagnostic functions and their names beyond generic 'posterior analysis, model checking, comparison'
- Whether the package is actively used by the broader Bayesian modeling community or primarily internal to ArviZ ecosystem
- Performance characteristics or scalability limits for large posterior samples

## Package facts

- License: not declared (permissive)
- Python support: supports_current
- Install friction: low
- Maintenance: active
- Downloads: 730.9K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

bayesian model diagnostics, posterior analysis statistics, bayesian inference diagnostics, model comparison metrics, bayesian convergence diagnostics, posterior predictive checking, bayesian model validation, bayesian-inference, diagnostics, statistics

[View on SkillFed](https://skillfed.io/packages/arviz-stats) · [View on PyPI](https://pypi.org/project/arviz-stats/)