A new model for determining the traffic accident black spots using GIS-aided spatial statistical methods

Traffic accidents are one of the important problems in our country as it in the world. The World Health Organization case reports published in 2015 stated that approximately 1.25 million people died each year and more than 50 million people injured as a result of traffic accidents in the world. Considering this situation, it is seen that traffic accidents are mostly human originated and one of the major problems that is negatively affecting life. In this context, many investments and many studies have been performed on the determination of traffic accident black spots to reduce traffic accidents. The current study aimed to get a descriptive model for determining the traffic accident black spots using model-based spatial statistical methods. These methods are Poisson regression, Negative Binomial regression and Empirical Bayesian method. The ultimate goal of this study was to build a model that allowed evaluating all the methods together in Geographic Information Systems (GIS) which is quite widely used nowadays. In the present study, the data were obtained from 300 thousand traffic accidents occurred on 2408 different state roads during the years from 2005 to 2013 from the General Directorate of Highways. The state roads of Turkey were divided into 32,107 sub-segments with the length of 1 km. Based on the study results, 126 sub-segments were determined as traffic accident black spots depending on the method used. According to comparison of the methods used in the present study, the Empirical Bayesian method provided the best results in terms of accuracy and consistency. (C) 2017 Elsevier Ltd. All rights reserved.