Abstract
The majority of bicycle crash studies aim at determining risk factors and estimating crash risks by employing statistics. Accordingly, the goal of this paper is to evaluate bicycle–motor vehicle crashes by using spatial and temporal approaches to statistical data. The spatial approach (a weighted kernel density estimation approach) preliminarily estimates crash risks at the macro level, thereby avoiding the expensive work of collecting traffic counts; meanwhile, the temporal approach (negative binomial regression approach) focuses on crash data that occurred on urban arterials and includes traffic exposure at the micro level. The crash risk and risk factors of arterial roads associated with bicycle facilities and road environments were assessed using a database built from field surveys and five government agencies. This study analysed 4120 geocoded bicycle crashes in the city of Antwerp (CA, Belgium). The data sets covered five years (2014 to 2018), including all bicycle–motorized vehicle (BMV) crashes from police reports. Urban arterials were highlighted as high-risk areas through the spatial approach. This was as expected given that, due to heavy traffic and limited road space, bicycle facilities on arterial roads face many design problems. Through spatial and temporal approaches, the environmental characteristics of bicycle crashes on arterial roads were analysed at the micro level. Finally, this paper provides an insight that can be used by both the geography and transport fields to improve cycling safety on urban arterial roads.
Highlights
As a key component of sustainable transportation systems, cycling has been actively promoted in cities throughout the world [1,2]; bicycle-related crashes have been associated with increasing numbers of fatalities and injuries [3,4,5,6] and the risk of crashes prevents people from using bicycles [7]
By using a two-stage strategy to assess bicycle crash risks, a spatio-temporal workflow opens new research directions for the analysis of traffic crashes (i.e., models aiming at estimating the risk of bicycle–motorized vehicle (BMV) crashes from a macro scale to a micro scale)
BMV crashes may be explained by a series of spatial and temporal phenomena
Summary
As a key component of sustainable transportation systems, cycling has been actively promoted in cities throughout the world [1,2]; bicycle-related crashes have been associated with increasing numbers of fatalities and injuries [3,4,5,6] and the risk of crashes prevents people from using bicycles [7]. Most studies neglect the regional impact of bicycle safety on the Sustainability 2019, 11, x FOR PEER REVIEW macro-scale level. This holds especially true for crash analyses of specific locations, e.g., intersections or other carraesah[-1p2,r1o3n],eanlodcaarteiolinkeslyonto tbheeomveirc-drios-psecraselde dleuvrienlg. Mndeatniwmheil.eT, fhroemreafogeroeg, rsaipghnicifialcpaenrstpgeactpivse,are present in currentmesbapinecycyiacslltleuydcairpeaspsliihgenssottrouedctrihaeessh.loacnaatliyosneasl influence of safety attributes on the micro-scale level This of geographic areas, such as spatial autocorrelation or many different kinds of cluster analyses on the macro-scale (regional) level. Despite limitations due to issues of privacy, a few studies have explored the influences of human and vehicle factors on bicycle collisions [24,25,26]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
