Evaluating Crash Risk in Urban Areas Based on Vehicle and Pedestrian Modeling

Itzhak Omer, Victoria Gitelman, Yodan Rofè, Yoav Lerman, Nir Kaplan, Etti Doveh

Research output: Contribution to journalArticlepeer-review

Abstract

This article presents a method for investigating the spatial distribution of vehicle and pedestrian traffic crashes relative to the volume of vehicle and pedestrian movement in urban areas. This method consists of two phases. First, vehicle and pedestrian traffic volumes on the street network are modeled using a space syntax configurational analysis of the network, land use data, and observed traffic data. Second, crash prediction models are fitted to the traffic crash data, using negative binomial regression models and based on traffic volume estimates and street segment lengths. The method was applied in two areas in Tel Aviv, Israel, which differ in their morphological and traffic characteristics. The case-studies illustrated the method's capability in identifying hazardous locations on the network and examining relative crash risks. The analysis shows that an increase in vehicle or pedestrian traffic volume tends to be associated with a decrease in relative crash risk. Moreover, the spatial patterns of relative crash risks are associated with the design characteristics of urban space: areas characterized by dense street networks encourage more walking, and are generally safer for pedestrians, while those with longer street segments encourage more driving, are less safe for pedestrians, but safer for vehicles.

Original languageEnglish
Pages (from-to)387-408
Number of pages22
JournalGeographical Analysis
Volume49
Issue number4
DOIs
StatePublished - Oct 2017

Fingerprint

Dive into the research topics of 'Evaluating Crash Risk in Urban Areas Based on Vehicle and Pedestrian Modeling'. Together they form a unique fingerprint.

Cite this