Burglary Prediction using Deep Learning

Residential burglary is still prevalent in most cities. It is sometimes difficult to predict where this kind of crime will happen. However, many cities have made their crime data available to the public. By analyzing these big crime data sets, it is possible to discover the patterns of urban structures that increase the risk of burglaries. In this study, deep learning was utilized to extract relationships between various house and environmental metrics and burglary. Through these relationships, the houses that have the higher risks of being burglarized can be identified. The City of Austin, Texas has been used for our case study since the city discloses various data sets including crime, street networks, demographics, and many others. This study can be used to deploy police patrols to the areas that are likely being burglarized. This can also provide an insight for making the urban environment safer by changing environmental cues and structures