While the intitial exploration of these three variables in regards to auto theft begins to describe the spatial and temporal aspects of auto theft in Vancouver, many other considerations may be pursued to increase the accuracy of these findings. Many of the potential additions to the work that was completed in this project that would increase viability of the results affect the spatial components of the study. Because both proximity to bus stops and zoning are spatial variables, some degree of colinearity is expected. Zoning would be expected to be connected to bus service, as bus routes are more likely to serve particular districts, such as commercial, and less likely to serve others, such as low density residential. In order to provide a more robust study, the degree of colinearity that exists between bus stop proximity and zoning type should be determined. A more general option that may be a possibility would be the procurement of more accurate data regarding the location of auto thefts. While this may be unfeasible as a result of the data being supressed, the accurate location of thefts may provide more insight into proximity relationships. The most obvious suggestion for further work on the subject of auto theft in Vancouver is the introduction of more variables that may influence auto theft, either directly or as a proxy. Finally, the creation of a predictive surface based on these variables would be a useful end product to visually describe the complex phenomenon of auto theft in Vancouer.