Masters candidate in Biostatistics, Dillon Corrigan, will present:

“multinomss: An R Package for Retrospective Space-Time Cluster Detection Using the Multinomial Scan Statistic”

Plan B Adviser: Mark Fiecas

Abstract: Spatial and space-time scan statistics are commonly used in epidemiology, veterinary population medicine, and other disciplines that rely on geospatial data to determine whether point processes are random or if clustering of points can be detected. To this end, we develop an R package, multinomss, that implements the multinomial space-time scan statistic and produces easy-to-read results and interactive visualizations. The Raptor Center (TRC) at the University of Minnesota College of Veterinary Medicine conducts research to identify emerging and persistent environmental issues related to raptor health. We will demonstrate the application of multinomss using TRC data and show equivalency to the results obtained via SaTScan, a popular stand-alone software that unfortunately is not open-source. Statistical and computational limitations of the multinomial space-time scan statistic will also be discussed.