Citation

Pasupuleti N, Pulugurtha SS. Procedia Soc. Behav. Sci. 2013; 104: 1188-1197.

Copyright

(Copyright © 2013, Elsevier Publishing)

DOI

10.1016/j.sbspro.2013.11.215

PMID

unavailable

Abstract

This paper focuses on examining the role spatial extent in capturing off- and on-network characteristics and developing macro-level intracity truck crash estimation models. A Geographic Information Systems (GIS) based method was developed to identify intracity truck crash zones, capture spatial data within 0.25-, 0.5- and 1-mile buffers for each selected zone, and examine relationships between demographic, socio-economic, land-use, and on-network characteristics and intracity truck crashes. Non-linear truck crash estimation models (Negative Binomial with log-link) were then developed using selected independent variables that are not correlated to each other. The working of the method and development of models is illustrated using 2008 data for the city of Charlotte, North Carolina located in the United States. Spatial characteristics extracted using 0.5-mile buffer width was observed to provide more statistically meaningful results to model intracity truck crashes. They were found to be positively correlated to industrial areas and areas with large residential lots but negatively correlated to densely populated areas. Zones with a median and inadequate turning radius seem to be more prone to intracity truck crashes. The method and findings from this research assists practitioners to proactively identify potential high risk areas, to understand the role of various factors on intracity truck crashes, and to develop plans to enhance safety.