The development of an Urban Crime Simulator has resulted in a software tool that allows users to simulate changes in crime rates of urban neighborhoods with his/her own GIS dataset, to select data attributes from the GIS dataset, to assign weights to the attributes based on his/her professional experience or knowledge acquired from literature. Urban Crime Simulator was developed using C#, a computer programming language as a part of Microsoft Visual Studio, and library routines in ArcEngine. Urban Crime Simulator groups urban neighborhoods based on users’ definition of neighborhood characteristics into clusters and uses the resulting clusters as the basis for estimating changes in crime rates. Users update individual urban neighborhoods with projected/simulated growth such for Urban Crime Simulator to estimate changes in crime by searching through clustered neighborhood for the neighborhood with the most similar profile. The crime rate of the neighborhood with the most similar profile as the updated neighborhood suggests an expected level of crime as the result of simulation.
Through this developmental effort, we argue that a combination of criminological theory, coupled with the concept of neighborhood life cycles, are a better approach to estimating/simulating changes in crime rates in urban neighborhoods. We recognize the limitation of using commercially available routine libraries as it requires royalty fees when distributing the resulting software tools. However, given the short time and limited budget, other options are limited. Finally, with the flexible modeling structure and flexible level of geographic aggregation that UCS accepts GIS data, we suggest that additional data mining tools be added to assist users of UCS better understand spatial and temporal trends of their urban neighborhoods before and during the simulation of urban crime.