Workshop on Bayesian Statistics

Estec, 18-20 December 2017


The Faculty organized a workshop on bayesian statistics.

It was given by Dr. Rafael S. de Souza, based on the warm recommendation of Eric Feigelson (who gave the statistics course which took place in 2015). Rafael has recently written a book on "Bayesian Models for Astrophysical Data” . You can find more information about Rafael here. Additional support/lectures to the workshop were provided by Dr. Emille Ishida.

The workshop was held on 18-20 December in Newton. The nominal attendance was here.

The intention was to spend some time on basics, i.e. the workshop was not primarily intended for experts. But also to make it very hands-on, with a good deal of time spent working out example problems.

It was recommended to prepare by reading parts of Rafael's book. During the workshop ample time will be given to practical exercise. Participants brought their laptops, having installed the packages listed here. Extra packages were installed on-the-fly in both R and python.


The initial agenda can be found here. The intention was to be flexible and adjust it according to the needs/wishes of the participants.

The actual programme that was implemented is shown below.

Day 1 - The Basics

Day 2 - Generalized Linear Models

  • Normal models
    Example in R using JAGS errors-in-measurements
    Examples in R using JAGS M-sigma relation

    Examples in Python

  • Time series AR(1) model
    Example in R using JAGS
    Example in Python

  • Gamma model
    Example in R

  • Lognormal model
    Example in R using JAGS- black body

  • Bernoulli model
    Example in R
    Fraction of red spirals in R using JAGS

  • Poisson model
    Example in R

  • Negative binomial model Population of globular Clusters in R using JAGS

  • Mixed models


Day 3 - Advanced topics