Using risk terrain modeling to forecast residential, commercial, and vehicle burglary in San Antonio, Texas

The purpose of this thesis was to apply Risk Terrain Modeling (RTM) to the crime of burglary in San Antonio, and in doing so, expand upon existing knowledge regarding its capacity as a spatial analysis tool. Burglary incident data and base maps provided by San Antonio Police were used to assess the effectiveness of RTM in forecasting residential, commercial, and vehicle burglary by determining where high levels of environmental risk tend to concentrate. Raster maps were created in ArcGIS, coded to represent the presence and magnitude of risk factors, and then layered to model a risk surface. The merged layers generated a composite risk value for each part of the mapped terrain, indicating where conditions were optimal for burglary (residential, commercial or vehicle) to occur.

Regression analysis was then used to test the predictive validity of each burglary model. Two inquiries provided the focus of this research: 1) whether RTM is effective in quantifying risk across two different landscapes of the city and 2) if RTM is applicable to residential, commercial, and vehicle burglary. Analysis and comparison of two demographically and socially diverse areas of the city revealed similarities and differences among risk factors for both sectors. Additionally, utilization of separate models for residential, commercial and vehicle burglary provided insight into how risk varies according to burglary subtype or category.