Exploratory spatial data analysis (ESDA) within a GIS to identify statistically significant clusters of childhood pedestrian collisions

Background/Purpose:

Motor vehicle pedestrian collisions are a significant problem for urban children. 12,206 pedestrian collisions involving children less than 19 years of age resulted in 421 fatalities in Hartford, Connecticut from 1997 through 2006. Geographic Information Systems (GISs) are increasingly used to analyze public health problems such as childhood pedestrian collisions. GIS is used to identify areas with high frequencies of collisions. Problem areas are then mapped with potential correlating factors. Despite increased use of GIS, important toolsets remain undescribed in the analysis of childhood pedestrian collisions. The purpose of this research is to demonstrate several important GIS methods that can be used to investigate childhood pedestrian collisions in Hartford, Connecticut.

Methods:

Exploratory spatial data analysis (ESDA) within a GIS was used to identify statistically significant clusters of childhood pedestrian collisions at two time periods: from 1986 through 1987 (n=234) and from 2005 through 2006 (n=90). Clusters found during each time period were then mapped to show possible correlation with several geographic, population-based, and ecological factors.

Results:

The findings highlight that GIS techniques can be used to show statistical evidence of the spatial patterns of childhood pedestrian collisions. Through this, we also show that the traditional methods used to compare rates across areas may be faulty, and can influence the results of regression analyses.

Conclusions:

The implications of this study are broad-reaching as the analysis simultaneously approaches the problem of childhood pedestrian collisions, while highlighting replicable GIS techniques that may be integral to a number of public health studies.