Course workshop. Topic: Introduction to Bayesian modelling of spatial and
Sujit K. Sahu, University of Southampton
This 1-day course will provide an introduction to Bayesian hierarchical modelling of spatial and spatio-temporal data. The course will cover
modelling and analysis of both point referenced data e.g. house prices, and areal level data e.g. county level cancer rates. Basic concepts of
spatial (or geo) statistics such as variogram, kriging, Matern correlation function, will be explained. These will be followed up
formal kriging based on Gaussian processes under a Bayesian hierarchical modelling set up. The methods will be extended to handle space-time
environmental monitoring data, such as air pollution, and will be illustrated using the spBayes and spTimer packages in R.
Conditional auto-regressive modelling for areal data will be considered and illustrated with several real life data sets. Bayesian versions of
these models will also be introduced with examples implemented in WinBugs.
If you are unsure about the suitability of your background for the course, please email Prof Sujit Sahu (S.K.Sahu@soton.ac.uk) who can advise.
Workshop outline timetable:
The workshop will run on July 15, 2014 starting at 9AM and finishing at 7PM with a break for lunch from 1PM to 3:30PM. Theory lectures will be
followed up by practical sessions where hands-on training using R will be provided. .
09:00 AM - 10:30 AM Introduction to Bayesian Modelling and MCMC
10:30 AM - 11:00 AM Break
11:00 AM - 01:00 PM Bayesian Modelling and Prediction of Point Referenced Spatial Data using spBayes
01:00 PM - 03:30 PM Lunch Break
03:30 PM - 05:00 PM Bayesian Modelling of Spatio-temporal Data using spTimer
05:00 PM - 05:30 PM Break
05:30 PM - 07:00 PM Bayesian Modelling of Areal Unit Data using WinBugs
The exact schedule may vary slightly from the above.
Post-graduate researchers including PhD students and early career researchers. Researchers from other disciplines but must be familiar
with practical Bayesian modelling using WinBugs or R. Knowledge of R is essential.