A spatio-temporal mapping to assess bicycle collision risks on high-risk areas (Bridges) - A case study from Taipei (Taiwan)

Abstract Most bicycle collision studies aim to identify contributing factors and calculate risks based on statistical data ( Loidl et al., 2016 ). The aim of this paper is to follow this approach, focusing on bicycle-motorized vehicle (BMV) collisions through a spatio-temporal workflow. For the spatial dimension (Kernel Density Estimation (KDE) method), a general estimation of the collision risks was obtained and the labour-intensive work of collecting counting data was avoided on the macro-scale level. The temporal dimension (negative binomial modeling method) focused on data from collisions occurring on bridges, enabling the inclusion of traffic exposure (counting data on the micro-scale level). Bridge collision risks and contributing factors related to road environment and cycling facilities were estimated using databases from eight government authorities and field investigation. For the presented case study, 2044 geo-coded bicycle collisions in the Taipei-Capital Region (Taiwan) were analysed. The data set covers three years (2015–2017) and includes all BMV collisions reported by the police. Through the spatial workflow, urban bridges were identified as areas with the highest density of collisions. This is unsurprising given that bicycle facilities on urban bridges face design difficulties due to limited space, discrepancy in elevation and traffic volume. Through this approach the characteristics of BMV collisions on bridges, traffic engineering, road environment, traffic control system, and driving behaviour were then analysed in the temporal dimension. This paper concludes by providing information relevant to traffic engineers concerning the enhancement of bicycle safety on high-risk areas in the city. Objectives In this paper, we aim to (1) understand the risk patterns of bicycle collisions spatially (where?) and temporally (when?), from a region-scale (macro) level to a location-scale (micro) level. To this end, a spatio-temporal (two-stage) workflow was developed for the exploration of the collision data. Through the spatial stage, urban bridges were identified as having the highest density of BMV collisions. Building on results from the spatial stage, we sought to (2) further explain “how” bridge infrastructure influences bicycle collisions in Taipei-Capital Region (Taiwan) by studying contributing risk factors. Countermeasures can thus be made to enhance bicycle safety. This paper is organized as follows. Section 2 provides an overview of the literature related to bicycle collision studies. Section 3 describes the study concept of the spatio-temporal workflow and its methodologies. Section 4 presents the study area and provides the descriptive statistics of BMV collisions. Section 5 describes the data used in spatio-temporal modeling. Section 6 discusses the main results and Section 7 finally concludes with the contribution and recommendation of this research.