applied statistics · Async
Bayesian Regression for Analysts
Priors without mysticism, posterior summaries stakeholders can read, and sensitivity checks that matter.
ZAR 1 480 · 7 weeks
Software: R · Method focus: Bayesian · Level: advanced · Assessment: notebook + brief
Description
Use Stan-style workflows conceptually with R interfaces. Focus on prior predictive checks, posterior intervals, and model comparison with WAIC-style discussion.
What is included
- Prior predictive simulation labs
- Posterior predictive checks with mentor feedback
- Sensitivity grid template
- Short readings on prior elicitation
- Compare Bayesian and frequentist intervals on same data
- Ethics note on double-dipping
Outcomes
- State priors in plain language tied to data scale
- Report credible intervals alongside practical significance
- Run a defensible sensitivity table
Dr. Jonas Weber
Computational statistician supporting economics RAs.
Reviews
Prior predictive section made our team agree on scales before touching Stan — rare alignment.
Credible vs confidence language module reduced confusion in our stakeholder deck.
Questions
We use open-source tools; you install packages locally.
We avoid measure theory; focus on computation and interpretation.
Hierarchical models with complex spatial structure are out of scope.