Citation

Schoner J, Tolford T, Putta T, Izadi M, Finfer R, Jatres D, Patterson D, Ruley J, Nigro J, Stickney R. Transp. Res. Rec. 2024; 2678(5): 542-555.

Copyright

(Copyright © 2024, Transportation Research Board, National Research Council, National Academy of Sciences USA, Publisher SAGE Publishing)

DOI

10.1177/03611981231189734

PMID

unavailable

Abstract

Vulnerable road user traffic deaths in the United States have increased in number and proportion over the last decade. This growing disparity points to a larger need to prioritize safety for vulnerable road users. Evaluating and predicting vulnerable road user crash risk is a data-intensive and complex process. This study aimed to make safety analysis easier and more accessible by (1) developing a modeling framework with minimal data input needs, (2) converting model outputs into cost equivalents to better link the results to project scoping processes, and (3) building this functionality into an online tool and dashboard. In this paper, we present an approach to modeling vulnerable road user crash risk that uses Bayesian probability updating and Markov chain Monte Carlo simulations to blend an existing published statistical model with simple roadway and crash data inputs, which we built into an online tool and dashboard called the Safer Streets Priority Finder. We applied the tool to crash data from the City of New Orleans and describe its application for roadway safety and transit planning use cases. Overall, in most contexts, we found that this modeling approach performed as well or better than sliding window analysis and traditional high injury networks, as it goes beyond just crash history, thus enabling it to estimate crash risk even when there is no history of crashes. This performance improvement, combined with ease of use, suggests the tool could improve on one of the most common safety analysis approaches used in field of transportation planning.

Language: en