CPD32 - Workshop: Model-based geostatistics: with applications in the
environmental and health sciences
17th and 18th November 2011
(9:00am - 5:00pm both days)
Flinders University, Adelaide
Presented by Professor Peter Diggle (CHICAS, Health and Medicine, Lancaster University)
Course description
This two-day course covers the core theory and methods of model-based geostatistics, as introduced by Diggle, Moyeed and Tawn (1998) and described in more detail in Diggle and Ribeiro (2007). The course also covers the implementation of these methods in the open-source R computing environment, using the geoR and geoRglm libraries. Theoretical ideas will be motivated by, and applied to, case-studies in environmental science and epidemiology.
The course assumes a working knowledge of statistical methods to the level of linear and logistic regression modeling. Familiarity with the R computing environment is desirable, but not essential. Course delivery will be a combination of lectures, software demonstrations and open-ended lab sessions where participants can gain experience of the geoR and geoRglm libraries with tutorial assistance on hand.
The course leader will also be available for informal discussion with participants during the open -ended sessions, and participants are welcome to bring with them open problems that they wish to discuss. However, it should be emphasised that this is an instructional course, not a workshop, and participants should not expect to be able to complete a substantive analysis of original data during the two days of the course.
Please note that the registration desk will be open from 8:30am on the first day of the workshop.
Outline programme
1. Introduction - motivating examples
This section uses examples from the environmental and health sciences to define the scientific scope of geostatistical methods and to motivate the methodological development of later sessions.
2. Linear models
The most widely used model in model-based geostatistics is the linear Gaussian model. In this section we define the model and show how it can be used to give a model-based rationale for the geostatistical methods for spatial interpolation and smoothing collectively known as (simple, ordinary or universal) kriging. We also discuss classical methods of parameter estimation and the use of simple transformations to extend the applicability of the model.
3. Bayesian inference
In this section, we cover the basic concepts of Bayeisan inference, apply these to the linear Gaussian model and compare Bayesian and classical methods through a case-study of spatial variation in rainfall in Switzerland.
4. Generalised linear models
In this section, we review classical generalised linear models for independently replicated data, develop their extension to geostatistical data and present an extended case-study of disease prevalence in equatorial Africa.
5. Geostatistical design
In this short section, we consider the core design issue of deciding where to place sampling locations within a spatial region of interest.
6. Geostatistics and marked point processes
In this section we discuss a current research topic in gesostatistical methodology. Conventional geostatistical methods assume that the process by which sampling locations are selected within a spatial region of interest is stochastically independent of the underlying spatial process of scientifc interest. When this is not so, a formally correct analysis treats the data as a marked point process.
The course draws heavily on material in Diggle and Ribeiro (2007).The mode of delivery will be a combination of lectures, software demonstrations and hands-on computing sessions. Participants are required to bring a laptop equipped with R package geoR. Click here to find out where to obtain this software.
Who should attend
The course assumes no prior knowledge of geostatistics or other spatial statistical methods but does require attendees to be comfortable with the formal language of stochastic processes and statistical inference, including: mean, variance and correlation functions; the likelihood function; linear and generalised linear models; classical and Bayesian inference.
The technical level of the most advanced material in the course is that of a first-year postgraduate statistics course. Numerate researchers in the environmental or health sciences should be able to appreciate the conceptual ideas presented, and to cope with the hands-on computing sessions.
About the Presenter
Peter Diggle is Distinguished University Professor of Statistics in the Faculty of Health and Medicine, Lancaster University, Adjunct Professor in the Department of Biostatistics, Johns Hopkins University, School of Public Health, and Adjunct Senior Researcher in the International Research Insitute for Climate and Society, Columbia University.
Between 1974 and 1983 Peter was a Lecturer, then Reader, in Statistics at the University of Newcastle upon Tyne. Between 1984 and 1988 he was Senior, then Principal, then Chief Research Scientist and Chief of the Division of Mathematics and Statistics at CSIRO, Australia.
Peter's research interests are in the development of statistical methods for spatial and longitudinal data analysis and their application in the biomedical, health and environmental sciences, with a particular focus on environmental and tropical disease epidemiolgy. He has published 10 books and around 200 articles on these topics in the open literature. He was awarded the Royal Statistical Society's Guy Medal in Silver in 1997, is a former editor of the Society's Journal, Series B and is a Fellow of the American Statistical Association.
Peter was founding co-editor, with his close friend and Johns Hopkins colleague Scott Zeger, of the journal "Biostatistics" between 1999 and 2009. He is a Trustee for Biometrika, and has served the UK Medical Research Council as a member of their Population and Systems Medicine Research Board.
Away from work, Peter plays mixed-doubles badminton with his partner Amanda and children Jono and Hannah. He is a keen cook, and should play more badminton than he does to counteract the effects of this. He also enjoys music, playing guitar and tenor recorder, and listening to jazz.
During his visit to Adelaide, Peter will also be presenting the E. A. Cornish Memorial Lecture for the SA Branch of SSAI, followed by dinner. Please see here for details.
Cost
Payment before 17 October 2011:
SSAI Members: $625 Non Members: $850 SSAI Student Members: $320 Non Member Students: $425
Payment after 17 October 2011:
SSAI Members: $675 Non Members: $900 SSAI Student Members: $370 Non Member Students: $475
Online registration is available here. To download a registration form, please click here.
Non-member students wishing to take advantage of the student rate need to use the hard-copy registration form and fax it together with their student id to SSAI, Fax No. 02 6251 0204.
Members need to be logged in with their username and password to take advantage of the member rates!
Venue
The venue is rooms 3.06 to 3.09 in the Health Sciences Lecture Theatre Complex (HSLTC) - building 53 on the following map (click here).
Travel Expenses
Occasionally workshops have to be cancelled due to a lack of subscription. Please contact the SSAI Office before making any travel arrangements to confirm that the workshop will go ahead, because the SSAI will not be held responsible for any travel or accommodation expenses incurred due to a workshop cancellation.
Cancellation Policy
Cancellations received prior to 10 November 2011 will be refunded in full. Cancellations need to be accompanied by a valid credit card number and expiry date which will be used to put the refund through. After 10 November 2011 no part of the registration fee will be refunded. However, registrations are transferable within the same organisation. Please advise any changes to eo@statsoc.org.au.
