Fast, beautiful and easy Bayesian modelling: Can you have it all?
summary
Needs fast reproducible fits for two-week release maps where a 25‑minute model time multiplies into days; uses Stan (C++) with HMC and NUTS for production and experiments with PyMC3 (pure Python, beta) for faster warm starts. Demonstrates a concrete hierarchical model on ~424 players and small-area cancer counts, implements a shrinkage estimator that pulls individuals toward a group mean, shows trace diagnostics and convergence issues for zero-observation units, and chooses Stan now for stability though PyMC3 promises speed.
AI-written summary · about
chapters (12)
- 0:00Fast, Beautiful and Easy Bayesian Modelling
- 4:29Motivating Bayesian analysis
- 4:57Population density in the US
- 5:50Batting rate in Baseball
- 6:17Who is the best Australian cricketer
- 7:01Unpooled
- 8:27Chris Lynn
- 12:18Binomial: A reminder
- 14:24Python
- 17:31Hierarchical model in Stan
- 18:41Posterior mu
- 22:39PyMC 3 outcomes
related talks
42mBayesian Non-parametric Models for Data Science using PyMC3
42mSolving Real-World Business Problems with Bayesian Modeling
01h 26mProbabilistic Programming and Bayesian Computing with PyMC
35mWhat we learned by running a large custom Bayesian...
01h 27mProbabilistic Python: An Introduction to Bayesian Modeling with PyMC
01h 27mBayesian Time Series
01h 30mAn introduction to Markov Chain Monte Carlo using PyMC3
40mHierarchical Bayesian Modelling with PyMC3 and PySTAN