Notes

The weekly instructional spine

These are the course’s weekly notes — the place to learn or re-learn the ideas if you missed class. Each week develops one stretch of the throughline, works concrete examples, names the mistakes that are easy to make, and maps that week’s reading. They are public, ungraded material; graded work lives in the LMS.

How to read a week

Every week note follows the same shape so you always know where to look:

Where this week starts — what last week left unfinished and what this week adds.
What you will be able to do — the week’s outcomes, stated as things you can perform.
Terms and notation worth fixing — the symbols and words as this course uses them.
Concept sections — the ideas, built rather than asserted, with nested development.
Worked example and Second worked example — complete cases carried through the arithmetic, the first with a shorter transfer of the same reasoning to a new setting, the second usually a diagnosis or a failure rather than a clean application.
The misreading to avoid — the wrong conclusion students actually draw, then dismantled.
Practice on your own — ungraded items for self-checking, with no worked responses.
Where to read more — the Bayes Rules! chapters this week aligns with, and the course pages.
Where this goes next — the bridge to the following week.

Several recurring examples run across the term so you can watch one problem develop rather than meeting a new one every week: a screening test at a walk-in clinic (Week 2), a seed lot’s germination rate that grows from a discrete updating table into a full Beta-Binomial model (Weeks 3–5, revisited in Weeks 8 and 12), a help desk’s call rate and a commute time for counts and means (Week 6, reused in Week 7), a municipal water utility carried through the regression, model-checking and decision weeks (Weeks 9–11, 14–15), and eight neighbourhood zones for hierarchical shrinkage (Week 13).

Software setup for R, VS Code, and Quarto is in Week 1.

The five parts

Part I — Foundations (Wk 1–2). What it means to reason with uncertainty, and discrete Bayes’ rule through diagnostic testing.

Part II — Building Bayesian models (Wk 3–7). From prior/likelihood/posterior to the Beta-Binomial model, prior sensitivity, posterior prediction, and simulation.

Part III — Synthesis & midterm (Wk 8).

Midterm synthesis

Part IV — Regression & model checking (Wk 9–12).

Part V — Hierarchy, decisions & synthesis (Wk 13–15).