Statistical analysis of spatial point patterns

Abstract

This project aims to test the broken windows theory of crime by examining the relationship between 311 calls for service and crime in New York through spatial analysis techniques. Using the inhomogeneous cross K-function, we found that 311 calls and crime were spatially clustered at inter-point distances of $100 - 850$ m. Controlling for the effect of spatial autocorrelation and a set of common socioeconomic indicators using spatial regression models, it was found that the volume of 311 calls is positively associated with crime, based on collected data for the time period from 2013 - 2017. Regional effects of the variables were investigated using a Geographically Weighted Regression model. The results suggest that 311 calls for service and crime are driven by common social processes and supports the broken windows theory.