Spatial point analysis of road crashes in Shanghai: A GIS-based network kernel density method

Becky P.Y. Loo, Shenjun Yao*, Jianping Wu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

46 Scopus citations

Abstract

As road crashes are constrained to a one-dimensional space, this paper analyzes the spatial distribution of road crashes with a GIS-based network-constrained kernel density method. A dissolving procedure is introduced before road segmentation, which can significantly reduce the undesirable effects during the segmentation process. The result of the sensitivity analysis reflects that the bandwidth imposes great impacts on the spatial distribution of density estimates. Different bandwidths may be considered for different types of traffic crashes. In particular, vehicle-pedestrian crashes in downtown areas tend to be highly localized and a narrower bandwidth is more appropriate. Vehicle-vehicle crashes at the suburb and rural areas, however, tend to happen in a less concentrated manner along a continuous stretch of dangerous road segments; and a wider bandwidth is more powerful in identifying these hot zones. Based on our results, administrations can gain more information on hazardous road locations, conduct investigations and propose improvement measures.

Original languageEnglish
Title of host publicationProceedings - 2011 19th International Conference on Geoinformatics, Geoinformatics 2011
DOIs
StatePublished - 2011
Event2011 19th International Conference on Geoinformatics, Geoinformatics 2011 - Shanghai, China
Duration: 24 Jun 201126 Jun 2011

Publication series

NameProceedings - 2011 19th International Conference on Geoinformatics, Geoinformatics 2011

Conference

Conference2011 19th International Conference on Geoinformatics, Geoinformatics 2011
Country/TerritoryChina
CityShanghai
Period24/06/1126/06/11

Keywords

  • GIS
  • crashes
  • kernel density
  • network
  • trafffic

Fingerprint

Dive into the research topics of 'Spatial point analysis of road crashes in Shanghai: A GIS-based network kernel density method'. Together they form a unique fingerprint.

Cite this