Geographic Information Systems has a range of applications, including research at UCCS.
Earlier this year, assistant geography professor Diep Dao co-engineered an analytical framework that uses GIS to track motor vehicle theft and other crime. The new system is called CrimeScape and may have significant implications for tracking crime.
Dao worked and published research with Jean-Claude Thill, a distinguished professor of geography at the University of North Carolina in Charlotte.
“We are trying to develop the methods, but at the same time [we are] trying to apply our methods into many different domains of application. We work with crime data in general, and we work with motor vehicle theft in particular,” Dao said.
Their research “focuses on advancing the traditional association rule mining (ARM) approach to capture the rich and multidimensional context that is anticipated to be associated with residential Motor Vehicle Theft (MVT) across urban environments,” according to the paper’s abstract.
In short, Dao and Thill sought to create a system that would consider factors that they believed were missing from current crime tracking research.
CrimeScape correctly produced some results that aligned with pre-stated notions about crime tracking, such as the fact that neighborhoods with certain socioeconomic factors tend to have higher rates of MVT. Areas with low rates of higher education, low rates of homeownership and low income, for example, produced higher rates of MVT.
This result showed that the framework could produce results in-line with extant findings. However, the research also produced results that clashed with the current understanding of crime distributions.
Some residential areas, such as minority communities, are singled out for high crime rates. CrimeScape identified that many of these areas were mostly free from MVT.
Why? These areas were characterized by “low income but low business activity, high homeownership, low multiple-unit housing, low young male population, average percentage of population with education of high school and above average employment,” among other factors, according to the publication.
Both researchers, having a background in Geospatial Data and Geography, realized that current research lattices tend to ignore geographical factors because of the system they use, ARM.
Association rule mining (ARM) is a tool that finds possible correlations and relationships among large sets of data items, usually at random. In the case of crime mapping, the machine looks at large chunks of crime data and tries to find connections between criminals and crimes, in theory.
Dao and Thill took this system and experimented by accounting for location, and effectively combining social science with geospatial technology. This novel research application of ARM granted the fascinating results.
CrimeScape “studies associations between urban residential MVT on the one hand, and the fundamental spatial demographic fabric of urban neighborhoods of a city including social economic status, residential stability, and racial heterogeneity, on the other hand,” according to Dao and Thill’s paper.
CrimeScape drew from several data sources to accomplish its task, ranging from residential MVT incidences, to education and employment, and even proximity to shopping malls and Walmarts. They conducted their research in Charlotte, North Carolina.
Dao is now looking ahead in her research by continuing to improve and record results from CrimeScape and delving into new research potential by looking at the way large cities utilize indoor 3D space.
“The third dimension is particularly hard to visualize for indoor space in terms of how people use that space. Another side of my research is visualizing 3D indoor [spaces], looking at accessibility and urban planning for large cities,” Dao said.
UCCS offers a GIS certificate in their geography department. The certificate takes students through several integral courses around the field, such as Geovisualization (GES 4070) and Spatial Database (GES 4040).