Determination of accident black spots by using network screening and K-means clustering methods: a case study on D100 highway in Istanbul

Abstract

Traffic accidents cause fatality, injury and property damage, as well as traffic congestion, road safety, noise, air pollution, and so on. According to Global Status Report on Road Safety published by World Health Organization (WHO) in 2015, deaths after traffic injuries are predicted to become the seventh leading cause of death by 2030. Besides, economic losses are occurred in large quantities in the developing economies like our country, due to traffic accidents. Various precautions are taken to reduce traffic accidents, such as increasing the number of traffic controls, eliminating road-related defects, increasing the frequency of vehicle inspections, and developing accident prevention mechanisms. In order to reduce the severity and the number of accidents, it has importance to identify and improve the zones called accident black spots where traffic accidents have occurred frequently. In this study, it is aimed to determine accident black spots by using Geographical Information Systems (GIS) and network screening methods and reveal the positive and negative aspects of these methods. For this purpose, the intersections and road segments at Istanbul D100 highway were sorted separately using simple ranking, sliding window and peak searching methods and accident black spots are determined using the K-means clustering method.