Spatial statistics have been used in the study of lip cancer, disease epidemics, the effects of the environment on health, in detecting clustering of rare diseases, and in modeling the growth of cancerous tumors. Medical researchers in many areas of public health and medicine often use and work with maps and associated spatial statistics. In spite of a growing need, there are currently no software packages providing a comprehensive suite of easy to use spatial statistics functions. We propose to develop SPACE-STAT, an interactive spatial statistics module for the S-Plus data analysis and graphics environment with an emphasis on techniques and methods of interest to the medical community. The combination of SPACE-STAT and DataViewer, StatSci's advanced environment for the visualization, exploration, and analysis of spatial data, will provide powerful tools for medical researchers in their analysis of spatial data. The SPACE-STAT module will include exploratory and graphical methods, multiple regression with spatially correlated errors, and methods for analyzing spatial point patterns. Also included are more specialized methods such as the auto-Poisson models, general linear mixed models, models for extra-Poisson variation, models and tests for space-time clustering, spatial models of tumor growth, and robustified versions when available.
