Taylor & Francis Group
Browse

Identification of contributing factors and hotspots of pedestrian-vehicle collisions in urban areas: An alternative hybrid method for small cities

Download (315.95 kB)
journal contribution
posted on 2022-08-01, 16:40 authored by Letícia Oestreich, Brenda Medeiros Pereira, Alejandro Ruiz-Padillo

Traffic safety numbers in developing countries show a serious global problem that requires studies and actions to promote safety for the most vulnerable users, such as pedestrians, as a way of contributing to sustainable urban mobility. Therefore, this research proposes a hybrid method to identify contributing factors and critical points according to the profile of crashes involving pedestrians being run over by motor vehicles in small cities.

The proposed method uses multivariate data analysis, specifically Exploratory Factor Analysis (EFA) and Cluster Analysis (CA), combined with geospatial tools. The study scenario corresponds to pedestrian-vehicle collisions that occurred in the urban area of Cachoeira do Sul, a small city in Rio Grande do Sul, Brazil, recorded in police reports between 2015 and 2018. Additionally, Geographic Information Tools (GIS) were used to collect characteristics of the road environment.

According to EFA, a set of attributes configured as contributing factors are discussed according to the components of the traffic system: user, road, vehicle, and environment. Subsequently, a CA was applied with the road-environmental factors, and three pedestrian-vehicle collision profiles were extracted, with clustering according to road segments with homogeneous characteristics. Moreover, a Kernel Density Estimate (KDE) was used to obtain heat maps of the pedestrian collisions profiles. The results allowed interventions to be proposed according to the contributing factors of each profile.

The multivariate statistical analysis supported the discussion of the factors contributing to the occurrence of pedestrian-vehicle collisions in small cities in developing countries. The spatial visualization of the results showed urban corridors with a high risk for pedestrians and the identification of hotspots in the city. In addition, the study demonstrated that the hybrid method tested is an alternative for the government of cities with few resources to promote safety mobility.

Funding

This work was supported by CAPES (postgraduate research grant number 88887.522083/2020-00) and by CNpq (research grants numbers 308870/2018-2, 135627/2021-4 and 422635/2018-9).

History