This study aims to compare and quantify the impact of traffic volume on hotspot identification. The data consist of geometric and traffic features of freeways of California for the six-year period (2005–2010). Five functional roadway classifications were used for analyzing the role of traffic volume in different environments. Four hotspot identification (HSID) methods were selected for this study, namely, Empirical Bayesian count with volume (EBWT), Empirical Bayesian count without volume (EBWOT), crash rate (CR), and crash number (CF). To determine the superiority of the above methods, four evaluation tests were conducted which include Site Consistency Test (SCT), Method Consistency Test (MCT), Total Rank Difference Test (TRDT), and Total Performance Difference Test (TPDT). The safety performance functions (SPF) that include the traffic volume show a better fit to crash count than do the ones without traffic volume.

The results also show that EBWT mostly comes out to be the superior method as indicated by the four tests, followed by EBWOT, CF, and the worst performer CR. The advantages associated with the inclusion of traffic volume in SPFs are also transferred to the HSID with EBWT showing the best performance in most cases.

Cheng, W., Gill G.S., Loera, L., Wang, X., & Wang, J.H. (2017). Evaluation of the impact of traffic volume on site ranking. Journal of Transportation Safety & Security, 10(5), 491-505.