Geospatial data integration and modeling for the investigation of urban neighborhood crime

Abstract

With more than half a century of development, Geographic Information Science (GIS) has evolved to become an interdisciplinary field of spatial thinking, geographic knowledge, geospatial technologies, and application practices. However, the integration of GIS in social science research is yet fully developed and more proactive emperical research examples are needed to continuously advance the application of GIS in social science research. To meet this challenge, in this article, the authors use the investigation of urban neighborhood crime as an experiment to examine the capability of geospatial technologies in the investigation of neighborhood crime in Oakland, CA, United States. First, a comprehensive theoretical framework is constructed with major neighborhood criminology theories to guide the empirical experiment. Second, a GIS-based methodological framework integrates geospatial data collection, integration, processing, and modeling on the one hand and advanced statistical methods on the other, to lead a data-driven examination of neighborhood crime. Specifically, a Random Neighborhood Sampling Matrix enables the generation of Hierarchical Adjustable Spatial Neighborhoods (HASNs). Areal Interpolation Matrixes allow the transformation of raw data in various geographic units to that in the HASN unit. Furthermore, a Neighborhood Accessibility Matrix accommodates the modeling of accessibility to nearest location-based crime factors from sampled neighborhoods. Third, multivariate statistics and multiple regression statistics are used to examine the relations between different types of neighborhood crime and their explanatory factors. Research rsesults indicate that the GIS-based methodological framework generates research findings highly consistent with those reported in the literature.