This one-week course teaches the applied Bayesian workflow for epidemiologic data using Stan. It is aimed at people who need to fit hierarchical models to real data and want a hands-on workflow rather than a theory course. The course condenses the Bayesian half of the Statistical Modeling of Infectious Disease Dynamics course and leans on the site’s Stan pages.

Draft proposal. This is an early sketch of a proposed short course. The course number, dates, fees, and daily schedule are placeholders and will be settled before the course is first offered.


Overview

IDE xxx (proposed) is a non-credit, one-week intensive course in Bayesian modeling with Stan. Participants work through the full Bayesian workflow: setting priors, checking them, writing and fitting a model, diagnosing the sampler, and checking the fit against data. Labs use cmdstanr in R or cmdstanpy in Python, and the week ends by applying the workflow to a transmission model.

By the end of the course, participants will be able to:

Who should apply

The course is aimed at graduate students, postdocs, and researchers who fit models to epidemiologic data.

Prerequisites. Working knowledge of R or Python is expected, along with a basic grounding in regression. No prior Stan or Bayesian experience is assumed.

Format and delivery

Course content and topics

Reproducibility conventions

Labs follow the reproducibility habits used across IDEEEP work. Every stochastic run is seeded (the course uses a shared seed so results match across machines), and lab code passes the seed to Stan explicitly rather than relying on defaults. When the sampler reports divergent transitions, the standard first step is to raise adapt_delta (for example from the default of 0.8 toward 0.95 or higher) and, where needed, to reparameterize the model. Participants record the seed, adapt_delta, and other sampler settings alongside their results so runs can be reproduced.

Day-by-day timetable

DayMorning lectureAfternoon lab (Stan)
1The Bayesian workflow; priorsPrior predictive checks
2Writing and fitting models in StanA first Stan model with cmdstanr / cmdstanpy
3MCMC diagnosticsReading diagnostics; setting adapt_delta
4Hierarchical and partial-pooling modelsFitting a hierarchical model
5Posterior predictive checks; transmission modelsFitting and checking a transmission model

Site resources

The course draws on material already published on this site. Participants can read ahead or review afterward.

Two further topics, prior predictive checks and posterior predictive checks, are planned as site pages and will be linked here once published.

See Programs for how this short course fits alongside the degree tracks and other offerings.

Fees and how to apply

Fees. @placeholder

How to apply. @placeholder

AI and academic integrity

Large language models such as ChatGPT, Claude, and Gemini can support your learning, and you are welcome to use them. If you do, cite the tools you used and describe how you used them. These tools do not replace your own understanding of the material, and you remain responsible for the accuracy of your work and any citations. Using them without attribution is plagiarism.

Proposal change notice

This is a draft proposal. Its content, structure, dates, and fees are subject to change before the course is offered.