Stan Workshop for BES conference in December 2017

Required packages

To make the most of the coding session please come to the workshop with the following packages installed:

• rstan (v2.16 or higher)
• bayesplot (v1.4 or higher)
• brms (v1.10 or higher)
• ggplot2
• reshape2
• plyr
• MASS

In addition windows users will most certainly need to install RTools, please follow the instruction given on this page.

To ensure a smooth start in the coding session check if you can run Stan models from R by running:

library(rstanarm)
stan_glm(mpg ~ cyl,family="gaussian",data=mtcars)

Structure of the workshop

• 45min general introduction into Bayesian data analysis, 15min Lionel, 15min Maxime
• 45min interactive coding session using brms and rstanarm to explore typical Bayesian Data Analysis workflow (fit models, checking, model expansion, setting priors, inference). The following models will be fitted:
  • Linear model
  • Generalized Linear Model (Negative Binomial)
  • (Mixed-effect models, varying intercept, varying intercept and slopes, crossed and nested random effects)
  • (Zero-inflated overdispersed poisson model)