Sunday, 5 July 2020 | 9:00 - 17:00 | Location: TBDInstructors:
- Peter Diggle
- Emanuele Giorgi
Summary:
The core objective of geostatistics (Diggle et al., 1998) is to make inference on a spatially continuous surface using sparsely sampled data from a geographical region of interest. Geostatistics was originally developed to construct predictive maps of the likely yield from potential mining operations, but is now used in many different scientific disciplines, including environmental monitoring, agricultural science and epidemiology. In this short course, we will introduce the underpinning principles of geostatistics and show their application to current global health issues, from mapping of infectious diseases in Africa to pollution monitoring in the United States.
Topics of the course are: 1) linear geostatistical models; 2) geostatistical models for disease mapping; 3) preferential sampling. These will be covered through a mix of lectures and lab sessions, keeping the mathematical formalism to the minimum required. Lab sessions will use the PrevMap package (Giorgi and Diggle, 2017) available from the Comprehensive R Archive Network, which implements state-of-the-art geostatistical methods using both classical and Bayesian methods of inference.
References:
Diggle, P.J., Moyeed, R.A. and Tawn, J.A. (1998). Model-based geostatistics. Applied Statistics, 47, 299-350
Giorgi, E., Diggle, P.J. (2017). PrevMap: An R package for prevalence mapping. Journal of Statistical Software 78:1-29.
Prerequisites:
1. Good knowledge of probability theory
2. Good knowledge of likelihood-based inference
3. Basic knowledge of generalised linear models (desired but not necessary)
4. Some knowledge of Poisson processes (desired but not necessary)
5. Basic skills in R programming (desired but not necessary)
Outline:
| 0900-1030 |
Geostatistical exploratory analysis - Brief historical introduction to geostatistics; the definition of geostatistical problems with examples; how to test for spatial correlation using the variogram (R). |
| 1030-1100 |
BREAK |
| 1100-1230 |
The class of linear geostatistical models - Linear geostatistical models and methods of inference using both classical and Bayesian approaches; Applications to mapping of malnutrition of Ghana (R). |
| 1230-1400 |
LUNCH |
| 1400-1530 |
Geostatistical models for disease mapping - The class of generalized linear geostatistical models with focus on Binomial and Poisson geostatistical models; how to test spatial correlation using counts data (R); Applications to mapping of malaria prevalence in Kenya and density of Anopheles mosquitoes in Cameroon (R). |
| 1530-1600 |
BREAK |
| 1600-1730 |
Introduction to preferential sampling - Review of point processes with focus on Poisson and log-Gaussian Cox processes (LGCPs); definition of weakly and strongly preferential sampling schemes with examples; methods of inference for LGCPs using Monte Carlo maximum likelihood; applications to lead pollution monitoring in Galicia, Spain, and ozone concentration mapping in Easter United States (R). |
NOTE: Lab sessions are labelled at (R). All of the data-sets used in the course are freely downloadable at the following link: sites.google.com/site/mbgglobalhealth/data. Each participant will be provided with slides, data-sets and R code beforehand.
Learning Outcomes:
On completion of the short course, the participants should be able:
to test spatial correlation using Monte Carlo methods based on the variogram;
to fit generalised linear geostatistical models using classical and Bayesian methods of inference;
to fit linear geostatistical models under preferential sampling;
to carry out spatial prediction for an outcome of interest;
to correctly interpret the results from a geostatistical model, from model fitting to spatial prediction.
About the Instructors:
Peter Diggle is Distinguished University Professor of Statistics in the Faculty of Health and Medicine, Lancaster University. He also holds honorary positions at the Johns Hopkins University School of Public Health, Columbia University International Research Institute for Climate and Society, and Yale University School of Public Health. His research involves the development of statistical methods for analyzing spatial and longitudinal data and their applications in the biomedical and health sciences.
Professor Diggle has developed and taught a wide range of courses in a university context, including undergraduate and postgraduate degree courses in probability and statistics, and statistics service teaching aimed at particular client groups. He has developed and taught short courses for CPD programmes in the UK and overseas, in topic areas related to spatial point processes, geostatistics, time series and longitudinal data. He has taught on the UK's EPSRC-funded national Academy for PhD Training in Statistics from its inception in 2008 until 2014.
Emanuele Giorgi is a Lecturer in Biostatistics and member of the CHICAS research group at Lancaster University, where he formerly obtained a PhD in Statistics and Epidemiology in 2015. His research interests involve the development of novel geostatistical methods for disease mapping, with a special focus on malaria and other tropical diseases. In 2018, Dr Giorgi was awarded the Royal Statistical Society Research Prize "for outstanding published contribution at the interface of statistics and epidemiology, spanning the development of spatial statistical methods, their application to a range of substantive problems in global population health research and their implementation in open-source software." He is the lead developer of PrevMap, an R package where all the methodology of the short course has been implemented. Dr Giorgi has delivered workshops and short courses on model-based geostatistics for a variety of non-statistical audiences, including veterinarians, ecologists, epidemiologists and clinicians in Belgium, Italy, the UK, Malawi, Mexico and Kenya. Since 2015, Dr Giorgi has been teaching courses in geostatistics and statistical modelling at the African Institute for Mathematical Sciences in Ghana, Tanzania and Cameroon. In 2018, Dr Giorgi has obtained a Postgraduate Certificate in Academic Practice which enabled him to become a fellow of the UK Higher Education Academy.
Recommendations for the Course:
Diggle P, Giorgi E (2019) Model-based Geostatistics for Global Public Health: Methods and Applications. CRC/Chapman & Hall.
Starting from February 2018, the book can be bought online at:
https://www.crcpress.com/Model-based-Geostatistics-for-Global-Public-Health-Methods-and-Applications/Diggle-Giorgi/p/book/9781138732353
CRC press has offered to provide promotion codes to the participants to obtain a 20% discount on online purchases.
The course requires all participants bring their own laptops with R and the PrevMap package installed.