Abstract

Automated Process of Identifying High Crash Clusters for Roadway Segments
Track: Transportation
Authors: LIXIN HUANG

Identifying the high crash clusters is critical to develop the high crash location list that can be used to prioritize the tasks of mitigating the number of crashes in the high crash roadway segments. The challenge to identify the crash clusters is selecting the appropriate bandwidth, which may affect the number of crash clusters in the roadway segment. GIS provides a spatial statistic tool called multi-distance spatial cluster analysis. This tool is based on the Ripley's K function and can analyze the spatial pattern of the incident point data. It can tell if the incident point data are spatially clustered or dispersed. The appropriate bandwidth can be derived from the aforementioned tool. In order to automate the whole process, the ModelBuilder application in GIS was used to model the workflow of identifying the high crash clusters for roadway segments.