You will need to be able to compile and run Stan models from R/Rstudio via either
cmdstanr. If you are not already an R user, we recommend using RStudio +
cmdstanr. A good guide to install RStudio is at https://moderndive.netlify.app/1-getting-started.html and installing cmdstanr is described at https://mc-stan.org/cmdstanr/articles/cmdstanr.html If you encounter any problems during installation, please ask for help at https://discourse.mc-stan.org/
To let you work quickly through the exercises, we recommend running our setup script that makes sure you have installed all you need and precompiles all models used in the exercises. Create an empty directory for the tutorial files and run
If you don’t want to use the setup script you’ll need to at least install the new SBC package. To install it, you’ll need the remotes package first:
Then you can install SBC via
If you are having technical problems running Stan models, visit on of the “Resolving Stan problems” tables in the social lounge.
Problems with tutorial tasks/questions should first be consulted with your colleagues at your table. If that doesn’t work for you for any reason, send a private message to one of the organizers (Martin Modrák, Hyunji Moon, Shinyoung Kim) - you need to click “People” in the right sidebar, find any of us and click “Send message”
If you want to discuss the results after the tutorial (e.g. wonder if you got the bonus questions right), feel welcome to ask at a dedicated thread on Stan forums.
Phil Clemson made a neat poster on using SBC for testing algorithms once we have a simulator and model we trust (unlike the tutorial which focuses on testing a model when we have an algorithm and a simulator we trust).