We have developed several methods for detecting geographic clusters of events–such as cases of disease or crime. These kinds of methods are used to identify statistical anomalies in geographical data that could point to a problem of concern–such as the presence of some toxin or disease causing organism.
Much of our work has been in adapting spatial scan methods methods for cluster detection. In one example, we worked on a method for detecting clusters that present in unusual and non-circular shapes. In another example, we introduced the idea of using quad trees to speed up the processes that searches for clusters. More recently, we developed a method for finding the range at which a location influences the risk of adverse events occurring, with applications in health, crime and environmental disaster research.
We continue to do work in this area with an emphasis on enhancing routine surveillance of events that occur at discrete spatial locations.