Bayesian Data Analysis course - Demos

BDA R demos

The BDA_R_demos repository contains some R demos and additional notes for the book Bayesian Data Analysis, 3rd ed by Gelman, Carlin, Stern, Dunson, Vehtari, and Rubin (BDA3).

Currently there are demos for BDA3 Chapters 2, 3, 4, 5, 6, 10, 11 and 12. Furthermore there are demos for RStan and RStanARM.

The initial demos were originally written for Matlab/Octave by Aki Vehtari and translated to R by Markus Paasiniemi. Recently more demos have been added for RStan, CmdStanR and RStanARM. Unless otherwise specified in specific files all code licensed under BSD-3 and all text, slides and figures licensed under CC-BY-NC 4.0.

See also Model Selection tutorial.

• ZIP file for all R demos
• Chapter 2
  • 2.1: Probability of a girl birth given placenta previa (BDA3 p. 37)
  • 2.2: Illustrate the effect of prior in binomial model
  • 2.3: Illustrate simulation based inference
  • 2.4: Illustrate grid and inverse-cdf sampling
• Chapter 3
  • 3.1_4: Normal model with unknown mean and variance (BDA3 section 3.2 on p. 64)
  • 3.5: Estimating the speed of light using normal model BDA3 p. 66
  • 3.6: Binomial regression and grid sampling with bioassay data (BDA3 p. 74-)
• Chapter 4
  • 4.1: Normal approximation for binomial regression model and Bioassay data
• Chapter 5
  • 5.1: Hierarchical model for Rats experiment (BDA3, p. 102)
  • 5.2: Hierarchical model for SAT-example data (BDA3, p. 102)
• Chapter 6
  • 6.1: Posterior predictive checking of normal model for light data
  • 6.2: Posterior predictive checking for independence in binomial trials
  • 6.3: Posterior predictive checking of normal model with poor test statistic
  • 6.4: Marginal posterior predictive checking with PIT test
• Chapter 7
  • See model selection tutorial
• Chapter 10
  • 10.1: Rejection sampling
  • 10.2: Importance sampling
  • 10.3: Importance sampling with normal distribution as a proposal for Bioassay model
• Chapter 11
  • 11.1: Gibbs sampling illustration
  • 11.2: Metropolis sampling + convergence illustration
  • 11.3_4: Metropolis sampling + convergence illustration
  • 11.5: Diagnostics with posterior and bayesplot packages
• Chapter 12
  • 12.1: Static Hamiltonian Monte Carlo illustration
  • 12.2: NUTS / Dynamic Hamiltonian Monte Carlo illustration
• CmdStanR, RStan and RStanARM
  • CmdStanR demos
  • RStan demos
  • RStanARM demos
  • Posterior predictive checking
  • Does brain mass predict how much mammals sleep in a day?
  • Non-linear model for traffic deaths in Finland
  • Extreme value analysis and user defined probability functions in Stan

BDA Python demos

BDA_py_demos repository some Python demos for the book Bayesian Data Analysis, 3rd ed by Gelman, Carlin, Stern, Dunson, Vehtari, and Rubin (BDA3).

Currently there are demos for BDA3 Chapters 2, 3, 4, 5, 6, 10 and 11. Furthermore, there are demos for PyStan.

Demos are in jupyter notebook (.ipynb) format. These can be directly previewed in GitHub without need to install or run anything.

Corresponding demos were originally written for Matlab/Octave by Aki Vehtari and translated to Python by Tuomas Sivula. Some improvements were contributed by Pellervo Ruponen and Lassi Meronen.

• ZIP file for all Python demos
• Chapter 2
  • 2.1: Probability of a girl birth given placenta previa (BDA3 p. 37)
  • 2.2: Illustrate the effect of prior in binomial model
  • 2.3: Illustrate simulation based inference
  • 2.4: Illustrate grid and inverse-cdf sampling
• Chapter 3
  • 3.1_4: Normal model with unknown mean and variance (BDA3 section 3.2 on p. 64)
  • 3.5: Estimating the speed of light using normal model BDA3 p. 66
  • 3.6: Binomial regression and grid sampling with bioassay data (BDA3 p. 74-)
• Chapter 4
  • 4.1: Normal approximation for binomial regression model and Bioassay data
• Chapter 5
  • 5.1: Hierarchical model for Rats experiment (BDA3, p. 102)
  • 5.2: Hierarchical model for SAT-example data (BDA3, p. 102)
• Chapter 6
  • 6.1: Posterior predictive checking of normal model for light data
  • 6.2: Posterior predictive checking for independence in binomial trials
  • 6.3: Posterior predictive checking of normal model with poor test statistic
  • 6.4: Marginal posterior predictive checking with PIT test
• Chapter 10
  • 10.1: Rejection sampling
  • 10.2: Importance sampling
  • 10.3: Importance sampling with normal distribution as a proposal for Bioassay model
• Chapter 11
  • 11.1: Gibbs sampling illustration
  • 11.2: Metropolis sampling + convergence illustration
• PyStan
  • PyStan test installation
  • PyStan demos

Python requirements

• python 3
• ipython
• numpy
• scipy
• matplotlib 2
• pandas (for some demos)
• pystan (for some demos)