Spatial Analysis for Crashes Involving 25- to 44- Year-Old Drivers in Kansas

Abstract

In this study, a descriptive and spatial analysis was conducted using crash data, 2013-2017, for 25- to 44- year-old drivers involved in fatal and serious injury crashes that occurred in the state of Kansas. The database integrates crash characteristics, concerning factors like environment, human, roadway, socio-demographic, and vehicle data for all the counties in the state to explore and understand the relationships between fatal and serious injury crashes and potential contributing factors. To understand the trend of fatal and serious injury crashes over the study period, a series of descriptive analyses were performed, which showed crash characteristics analyzed in this study that followed similar trends to most of the past crash studies. This study included 891 fatal and 2491 serious injury crashes along with 19 potential explanatory variables that were analyzed using Ordinary Least Squares (OLS) Regression to identify statistically significant factors, along with hotspots and outliers, in various geographical locations. A spatial model was created using two statistically significant variables to predict the total number of fatal and serious injury crashes, which was validated with the actual 2018 crash data. This spatial analysis and modeling were performed using ArcGIS Pro Software. Most of the statistically significant fatal and serious injury crash hotspots were located in and around Sedgwick county and northeastern Kansas counties like Douglas, Johnson, Leavenworth, and Wyandotte. The model validation for the crashes that occurred in 2018 was underestimated by eight percent, which is considered an acceptable variance. This study is the extension of the previous work that analyzed teenage drivers and also adopted a similar research methodology.