Instructor: Jesse Brunner
Lectures: Tu/Th 13:30 -- 14:45 in Spark 333 Labs: W 10:00 -- 13:00 in Eastlick G0098
Text: Statistical Rethinking, 2nd edition, by Richard McElreath
Tentative schedule (will be updated often, so check back)
| Week | Date | Topics | Readings (prior to class) & labs | Videos (Playlist) |
|---|---|---|---|---|
| Wk 1 | 7-Jan | Why simulating data is a super power; you are the captain of your ship <slides> | Ch 1. The Golem of Prague | |
| 8-Jan | R Shakedown & sampling distributions | Lab 1a Shakedown & Lab 1b Simulating data | 3B1B: Binomial distributions, Probabilities of probabilities | |
| 9-Jan | Draw your research: boxes and arrow, graphs, estimands <Slides> | (optional) Lundberg et al. 2021. What Is Your Estimand? Am. Sociological Rev. 86:532-565. | 3B1B Why “probability of 0” does not mean "impossible" | |
| Wk 2 | 14-Jan | Creating a generative model: conventions, specifying assumptions, and a workflow | Example data simulation (work along) | Lecture 1 |
| 15-Jan | Simulating data from your model (variables, constraints, distributions, and relationships) | Lab2_SimulatingYourOwnSystem.html / Lab2_SimulatingYourOwnSystem.qmd | ||
| 16-Jan | Reviewing the simulating process and why we do it. <slides> | Ch 2. Small Worlds and Large Worlds | 3B1B Bayes theorem, the geometry of changing beliefs & The medical test paradox and Redesigning Bayes' rule | |
| Wk 3 | 21-Jan | Bayesian models <slides> | Ch 2 (≥p28); Ch. 3. Sampling the Imaginary | Lecture 2 |
| 22-Jan | Running simple models & what to do with it: predicting from priors; prediction new data | Lab 3 Let's make, run, and check models! / Lab3_RunModels.qmd | ||
| 23-Jan | The levels of uncertainty, types of intervals <slides> | |||
| Wk 4 | 28-Jan | Gaussian and linear models as a good starting (ending?) point <slides> | Ch. 4. Geocentric Models (pp 71-110) | Lecture 3; 3B1B: But what is the Central Limit Theorem? |
| 29-Jan | Specifying a model; Simulating data from a model; | Lab 4 Linear models / Lab4_SimulatingFittingLinearModels.qmd | ||
| 30-Jan | Coding Example: estimating viral r | |||
| Wk 5 | 4-Feb | Coding Example: plotting |
Lecture 4 | |
| 5-Feb | Building tools to help with massive amounts of data; work on own data | Lab 5 Create some helpful functions | ||
| 6-Feb | continue to play with your own model | |||
| Wk 6 | 11-Feb | Quick overview of DAGs for causal inference <slides> | ||
| 12-Feb | Simulating different DAGs | Lab 6 Simulate from DAGs/Lab6_SimDAGS.qmd | ||
| 13-Feb | Drawing your own DAGs | |||
| Wk 7 | 18-Feb | Causal consequences | Ch. 5. The Many Variables & The Spurious Waffles | Lecture 5 |
| 19-Feb | Residuals, counterfactuals, categorical variables | Lab 7 Residuals, Counterfactuals, & categorical variables / Lab7_PostInference.qmd | ||
| 20-Feb | when causation is hidden | Ch. 6. The Haunted DAG & The Causal Terror | Lecture 6 | |
| Wk 8 | 25-Feb | Other aspects of causal inference | <What climate skeptics taught me about global warming> and <Hill 1965, The environment and disease: association or causation?> | |
| 26-Feb | Discussions of... | <Scott & scurvy>, & <Reality is weird> | ||
| 27-Feb | Entropy, KL distance, and Deviance <slides> | Ch. 7. Ulysses' Compass | Lecture 7 | |
| Wk 9 | 4-Mar | Predicting predictive accuracy | ||
| 5-Mar | Cross-validation and its short-cuts, regularization and fat tails | Lab 8 Prediction, regularization, & robust regression / Lab8_PredRegRobust.qmd | ||
| 6-Mar | Jesse gone... discuss amongst yourselves | |||
| Wk 10 | 11--13-Mar | Spring Break | ||
| Wk 11 | 18-Mar | Adding interactions | Ch. 8. Conditional Manatees | |
| 19-Mar | Lab 9 Interactions / Lab9_Interactions.qmd | |||
| 20-Mar | Marbles rolling around the posterior (HMC) <slides> | Ch. 9. Markov Chain Monte Carlo | ||
| Wk 12 | 25-Mar | How to use HMC & what to watch out for | Lecture 8; Optional: <Intro to Random Walk Metropolis algorithm>, <Metropolis-Hastings for constrained parameter>, <Intuition behind the Hamiltonian Monte Carlo> | |
| 26-Mar | Using MCMC for fun and profit | Lab 10 MCMC / Lab10_MCMC.qmd | ||
| 27-Mar | Why we worry about what's under the hood; problems and solutions | <Ben Lambert: The intuition behind the Hamiltonian Monte Carlo algorithm>, <Divergent transitions and what to do about them> | ||
| Wk 13 | 1-Apr | Other distributions, transformations, etc. | Ch. 10. Generalized Linear Model (p312-321) | |
| 2-Apr | Logistic regression | Lab 11 Distributions / Lab11_Distributions.qmd | ||
| 3-Apr | Poisson regression | Ch. 11. God Spiked the Integers (pp 345-358) | ||
| Wk 14 | 8-Apr | Fudging over multiple processes producing your data | Ch. 12. Monsters and Mixtures (pp 369-380) | |
| 9-Apr | Continuous mixture models | Lab 12 Distributions / Lab12_ContinuousMixtures.qmd | ||
| 10-Apr | Q&A: Getting comfortable in transformed space, back-transforms, etc. | |||
| Wk 15 | 15-Apr | Partial pooling in concept and practice | Ch. 13. Models With Memory | Lecture 12 |
| 16-Apr | Partial pooling and non-centered formulations | Lab 13 Distributions / Lab13_ModelsMemory.qmd | ||
| 17-Apr | Multilevel models in your own system | Lecture 13 | ||
| Wk 16 | 22-Apr | Correlations among parameters within clusters <slides> | Ch. 14. Adventures in Covariance (pp 435-454) | Lecture 14 |
| 23-Apr | Working with covariances | Lab 14 Covariances / Lab14_Covariance.qmd | ||
| 24-Apr | Continuous correlation structure <slides> | Ch. 14. Adventures in Covariance (pp 467-477) | Lecture 16 | |
| Intro to missing data | Ch. 15. Missing data & other opportunities (pp 491-512, 516-512 | Lecture 17 & Lecture 19 |