There are more than 5 million motor vehicle accidents (MVCs) in the United States every year. While some contributing factors are out of the hands of policy makers, others are not. In our investigation, we seek to better understand the relationship between the built environment of a community and the number and severity of crashes in that area. Features of the built environment (which include roads, crosswalks, bike paths, and traffic signals) can be augmented or replaced. To more easily discuss the number and severity of crashes in a small region, we define the concept of "crash density" as (severity * number of crashes)/ (population density) for a given area. By modeling the relationship between built environment features and crash density, we can better anticipate the number of severe MVCs a community may expect in certain areas