Abstract
Chicago and other major cities continue to get a lot of attention as they struggle with high crime rates. Chicago is currently ranked as one of the most crime-affected cities in the U.S and has witnessed a 50% increase in its murder rate in a year period. Law enforcement agencies usually face a complex problem when trying to efficiently assign resources to minimize crime overtime and across different city locations. The following project aims to review different data mining methods to effectively visualize crime trends and locations. Additionally, taking into account that crime analysis and prediction is a delicate subject, this report also intends to promote a discussion about community engagement and other crime reduction measures.
Crime analysis includes looking at data from two important dimensions: space and time. Space involves observing the characteristics of a particular region and surrounding areas. On the other hand, time involves observing and analyzing an event overtime. Using real crime data for the city of Chicago and several data mining models, we will explore different scenarios to provide options to address a real-world problem.
In this project, we will analyze Chicago crime data to understand different crime trends and identify relevant hot spots. By analyzing several characteristics of particular regions and observing criminal incidents overtime, we will attempt to formulate a solution and we will provide several recommendations. For this purpose, several models were trained and then were tested to guide us through our effort of finding effective predictive policing strategies. There has been a lot of debate regarding this topic due to ethical concerns and the risk of targeting vulnerable areas. Further discussion will be provided in the conclusions.