Bayesian Inference for Fun and Profit

on YouTube ↗01h 20m

" summary

Worked examples use Major League baseball (first 45 at‑bats) and Women's World Cup match data (48 matches) to compare full Bayesian posteriors against MLE/point estimates. Estimates come from Stan using CmdStan/commandstanpy workflows; the hierarchical model gives league mean ~0.27 and shrinks Roberto Clemente’s 0.400 to ~0.32, but requires raising adapt_delta to avoid divergences. Soccer model uses SPI priors and posterior predictive checks to show very wide credible intervals and that uncertainty, not just point estimates, changes decisions.

AI-written summary · about

" chapters (8)

1. 0:00Bayesian inference with Stan
2. 1:05Hamiltonian Monte Carlo explained
3. 7:10Multi-level regression models
4. 23:47Baseball batting case study
5. 34:07Running Stan with cmdStanPy
6. 47:51Women's World Cup model
7. 1:08:15Evaluating Bayesian models
8. 1:15:30Q&A

" related talks

• 01h 27mProbabilistic Python: An Introduction to Bayesian Modeling with PyMCChris Fonnesbeck · PyData London 2022
• 56mSports Analytics: Big Data in the Big TimeWayne Winston · PASS Business Analytics Conference 2013
• 42mBayesDB: Query the Probable Implications of DataRichard Tibbetts · Strange Loop 2015
• 25mScaling Bayesian ExperimentationRobert J. Neal, Paul Cho · Scale By The Bay 2018
• 01h 24mBayesianism and Survival AnalysisJake Coltman, Jacob Goodwin · PyData London 2016
• 39mThe Duct Tape of Heroes: Bayes RuleVincent D Warmerdam · PyData Amsterdam 2016
• 34mFast, beautiful and easy Bayesian modelling: Can you have it all?Rhydwyn McGuire · PyCon Australia 2015
• 42mThe Duct Tape of Heroes Bayesian statisticsVincent D Warmerdam · PyData London 2016