Abstract
The mission of the Sacramento State Police Department is to protect the life, property, and peace of mind of the students, faculty, staff and visitors of this university. To provide security to the Sac State community, various security services are provided such as diligent patrol, quick reaction to service calls, timely broadcasting crime alerts, etc. Predictive policing is an approach that uses data analytics to forecast new criminal acts by studying data about previous crimes. The objective of this method is to make investigation efforts more effective and develop effective strategies that prevent crime. This method helps police to deploy resources more accurately in place and time as well as identify people that can be involved in the act of crime. In this project, we have investigated whether predictive policing is applies to a smaller region with less amount of data available, such as the Sac State community. The goal is to use data analytic techniques and web technologies to make our campus a safer place. By analyzing historical crime data (Jan. 2017 – Jan. 2019), we have identified a few crime patterns. With the limited amount of data available, we have researched on the feasibility of using different analytical models and concluded on what is feasible and what is not, with regards to predict future crimes. As the outcomes of this project, we have presented a two-step prediction model, which can predict the crime hotspot on campus for a future weekday, and the corresponding crime type for a specific hotspot. We have also implemented a web-based prototype to showcase visualization and prediction results.