Spatial-Temporal Campus Crime Pattern Mining From Historical Alert Messages

A recent slew of violence-related events have incited a flood of discussions revolving around issues such as gun control and domestic safety abroad. An issue that hits close to home is the concern of safety at urban universities. Studying the geo-temporal distribution of crime within and around a university setting is important for understanding crime type occurrence patterns. These patterns can be mined from alert messages posted by universities on various media outlets, such as email, Twitter, and Facebook. We believe that the knowledge inferred from this data can be a crucial factor in creating a safe environment to protect students, faculty members, and administration. The observed patterns can help devise more effective crime prevention practices within and around a university campus, such as the optimization of the deployment of law enforcement resources according to recognized temporal and location patterns or the modification of patrol routes of police officers. Additionally, the observed geo-temporal patterns may help establish joint crime prevention programs between a university and the city.This research project aims to develop a system that automatically collects crime-logged data from publicly available sources, organizes it for mining, and creates visual mining tools to explore the data. We use Google Maps to render the data geographically.