a. Keimyung University, 1095 Dalgubeol-daero, Daegu 704-701, Republic of Korea
b. KAIST, N-22 #206, Daehak-ro, Yuseong-gu, Daejeon 305-701, Republic of Korea
c. Seoul National University,1Gwanak-ro, Gwanak-gu, Seoul 151-744, Republic of Korea
d. Korea University,145 Anam-ro, Seongbuk-gu, Seoul 136-713, Republic of Korea
Identifying hazardous locations on highways is a fundamental step in safety improvement programs and projects since it provides decision makers with a basis for allocating budgets and other resources in a cost-effective manner. Extensive research has been conducted to identify such locations. However, most studies have ignored the spatial characteristics of crash occurrences and the relative significance of injury severity. In this study, we develop a procedure for identifying hazardous locations on expressways based on geographically weighted regression(GWR)and equivalent property damage only (EPDO). GWR is a spatial regression method that can reflect spatial dependency and heterogeneous relationships between crash occurrences and other explanatory variables. EPDO is a comprehensive measure of crash occurrences weighted by the level of injury severity. We apply this procedure to a case study in Gyeongbu Expressway in Korea. The findings from our case study show that the procedure can identify hazardous locations on roadways while reflecting crash frequency and injury severity simultaneously with the comprehensive measure.