Notebooks and data for StanCon 2023 Tutorial on BRMS

This tutorial works through the example dataset from the paper Visualization in Bayesian workflow:

Gabry, J. , Simpson, D. , Vehtari, A. , Betancourt, M. and Gelman, A. (2019),
Visualization in Bayesian workflow. J. R. Stat. Soc. A, 182: 389-402. doi:10.1111/rssa.12378

• Published JRSS version: https://rss.onlinelibrary.wiley.com/doi/full/10.1111/rssa.12378
• arXiv preprint: https://arxiv.org/pdf/1709.01449.pdf (includes Supplementary Materials in appendix)

Exploratory Data Analysis

Notebooks "eda_air" are visualizations of the dataset in subdirectory data.

This dataset is GIS data, which requires either

• The R package sf - https://r-spatial.github.io/sf/index.html
• The Python package GeoPandas - https://geopandas.org/en/v0.4.0/index.html

Fitting Models in BRMS

Notebooks "fit_air_brms" use BRMS to specify and fit the 3 models from the paper. To run these notebooks you must install the R packages BRMS and cmdstanr. The input data is in JSON format, and therefore doesn't require the sf library.