Abstract

Using GIS to Predict High Crash Segments in Highway
Track: Transportation
Authors: Lixin Huang

It is important for emergency agencies to predict high crash segments in the highway. With that information available, emergency officials can allocate the limited resources more efficiently. Multiple years of historical crash data were obtained. Those crash data were processed by grouping them based on the day of the week due to the different crash patterns among each day. Then, clustered crash segments in the highway were identified by utilizing cluster analysis tools for the different time periods of the day, such as 6:00 am to 9:00 am. Those crash segments were ranked by the number of crashes. Crash segments with higher ranking are used to predict the potential high crash segments in the highway for a certain time period for different day of the week. A basic web application was also developed to visualize the high crash segments on the map.