Co-location analysis of pedestrian accident attributes for Ankara.

Abstract

Even though accidents with pedestrians are more likely to have a severe or fatal outcome, most research concentrates on the vehicle part of the accidents. To make better policies and infrastructure decisions it is crucial to understand how and why such accidents happen. In this article, 2.588 pedestrian accidents from the year 2018 that resulted in injury or death in the city of Ankara, Türkiye are considered, and different attributes are analyzed co-location-wise. Three types of analysis will be undertaken for these accidents: temporal, natural environment, and pedestrian characteristics. The season of the year, weekend-weekday, and time of day will be explored for the temporal analysis. Visibility (daylight, twilight, and darkness) and rain are the natural surroundings of interest. The analyzed pedestrian characteristics are age group, gender, clothing colour, and nationality. Generally, the accident properties are evenly distributed and mostly the higher co-locations occur with themselves. Three important results stand out in this study. Firstly, the type of precaution appropriate to different time periods of the day should be applied, taking into account the locations that have the strongest co-location with themselves. Secondly, a seasonal approach should be adopted in determining places where pedestrian safety needs to be increased in the city. Lastly, pedestrian safety measures targeting especially both '0-14 and 65 + age groups' and 'foreign citizenship' should be prioritized in the identified locations. The results show that the co-location methodology is a good fit for analyzing pedestrian accidents and a wider use of this methodology for other accident types and in general might be beneficial.