Course Dates: Mon, July 15 to Wed, July 17
Welcome to the 2024 SISMID course on Spatial Statistics for Epidemiology and Public Health! Spatial methods are now used in many disciplines and play an important role in epidemiology and public health. This module gives an introduction to spatial methods. In particular, we will present methods for assessment of clustering, cluster detection, spatial regression, small area estimation, and disease mapping. Methods will be described for both point data (in which cases and non-cases (or a sample thereof) have an associated point location) and count data (in which the numbers of cases and non-cases in a set of geographical areas are available).
An introduction to Geographic Information Systems (GIS) will be provided. The important extension to space-time analysis will be described, which is crucial for the analysis of infectious disease data with a spatial component.
Many examples will be presented, with analysis carried out in the R programming environment.
This module assumes knowledge of the material in Module 1: Probability and Statistical Inference, though not necessarily from taking that module. Some prior knowledge of R would be helpful.
This SISMID module exclusively uses R. We recommend R 4.0 and above. You can download a recent version here.
We also recommend using RStudio as the main GUI interface. You can install the free version here.
Installation of R and RStudio should proceed smoothly on most operating systems. Detailed instructions can be found here.
The module makes use of other specific R packages as well. You can install them using the install.packages("PACKAGE NAME HERE") command. R will let you know what packages you do not have installed when you try to run a script in full.
You will need to download the INLA package to run the code in this module.
The INLA package performs approximate Bayesian inference for latent Gaussian models. Installing the package is a bit different than normal since it is not on CRAN, the central software repository for R packages. Detailed instructions for installing the stable/testing can be found here.
In general, you will need to
- Clone the module repository from GitHub to a filepath on your local desktop.
- Open the code in RStudio Desktop by opening the R Project file
2023-SISMID-Spatial-Epi.Rproj.
- Introduction (Waller)
- GIS, Mapping (Waller)
- Areal Data (EDA, clustering) (Chang)
- Disease Mapping (CARs, SAR) (Chang)
- Spatial regresssion + Spatial Coefficient (Waller)
- Gaussian Process (Chang)
- Point process (Waller)
- Multivariate Process (Chang)
- Spatial infectious disease (SIR), ecology (Waller)
- Space-time models (Chang)
Spatial-Temporal Statistics with R
Geospatial Health Data: Modeling and Visualization with R-INLA and Shiny
Advanced Spatial Modeling with Stochastic Partial Differential Equations Using R and INLA
If you have any questions feel free to contact the following course instructors and TA's, or send a message in the class Slack channel.
Instructors: Lance Waller, Howard Chang
Teaching Assistant: Natalie Olson