Identification of Traffic Accident Hotspots using Geographical Information System (GIS)

Abstract

Limiting the number andseverity of traffic accidents is one of the major goals of road traffic safety management.The alarming rate of road accidents globally emphasizes the importance of an effective traffic safety management system. Identification of accident hotspots is the first step towards implementation of efficient traffic safety management.Until the arrival of Geographical Information System (GIS),traffic accident analyses have been performed based ontraditional statistical methods alone. The advent of GIS-based techniques has led toimproved traffic accident analysis by employing spatial statistics,enabling engineers and researchers to account for variation in the spatial characteristics of hotspot locations in the analysis. This paper discusses the different spatial and statistical methods that are employedintraffic accident hotspots identification. An example application of Planar Kernel Density Estimation (PKDE)for hotspot identification is presented based on crash data for Des Moines city of Iowa state. The effect of varying bandwidths in creating density mapsis investigated and the optimum bandwidth to obtain distinct hotspots is identified as 500 m for the chosen study area.The paper also discusses the scope for future research in traffic accident hotspot analysis.