Bandwidth Selection of Kernel Density Estimation for GIS-based Crime Occurrence Map Visualization

摘要

With the recent development of the Internet and satellites, the scope of utilization of the Geographic Information System (GIS)-based technology has been increasing. GIS refers to a system that integrates and manages spatial data and attribute data for objects with geographical locations to provide various forms of information, such as maps, diagrams and drawings In this paper, the Seoul Metropolitan Police Agency’s crime statistics data occurred over a specific month period have been linked to visualize three types of maps including point-based map, area-based map and density-based map, and then bandwidth selection of Kernel Density Estimation(KDE) is explored for density-based map. It is expected that such a crime map visualization research will identify the distribution of spatial characteristics of crime and further combine them with urban-engineered, socio-economic, and demographic characteristics to be used to establish policing policies that consider regional characteristics.