©[2011]

This dissertation presents a systematic approach to identify the determinants of gun shooting and forecast the distribution of gun shooting. Using frameworks of opportunities theories and social disorganization theories, this study examines the influences of urban features, including public housing, bus stops, liquor stores and bars, and schools, and neighborhood characteristics, on variations of rates of gun shooting in Newark and Irvington, NJ. A crime forecasting model is built based on the results of the determinants analyses in order to predict future shooting distributions. This study introduces a spatial statistics method, Conditional Locational Interdependence (CLI), to identify crime generators. CLI overcomes the limitations of current common method that requires arbitrarily determining the polygon size. In addition, this study utilizes several data analyses to address the spatial interdependence of the data. Besides, to address the non-Gaussian and spatial dependence of the data, poisson regression and negative binomial regression are employed and compared by adding spatial lag as one of the predictors. Further, smoothing method is used to convert census data to smaller units. Finally, risk terrain modeling is used to build a forecastin