Traditionally, incident management data have been collected for measuringand improving the performance of incident management. While these data are less detailed than crash records, they are timelier and contain useful data attributes that are typically not reported in the crash database. Thispaper proposed the Getis-Ord (Gi*) spatial statistics for identifying hotspots on freeways from incident management database while incorporating selected impact attributes into the analysis. The Gi* spatial statistics jointly evaluates the spatial dependency effect of the frequency and attribute values within the framework of conceptualized spatial relationship. Theapplication of the method was demonstrated through a case study using theincident database from Houston's Transportation Management Center (TranStar). The method successfully identified the clusters of high-impact accidents from over 30,000 accident records during 2006 to 2008. Although the accident duration was used as a proxy measure of its impact, the proposed method can be modified for identifying the locations with high-valued impacts using any other attributes provided that they are either continuous or categorical in nature and can provide meaningful implications. With the improvement of intelligent transportation system infrastructure and communication technology, the hot spot analyses for freeway network and arterials in the vicinity using incident management data have become a much more promising alternative. Freeway management agencies can use the hot spot results to provide visualized information to aid decision-making process in designing, evaluating, and managing incident management strategies and resources. The limitations of the method and possible future research are discussed in the closing.
Abstract