Network centralities and traffic safety in transportation networks: Evidence from the Amtrak railways

Abstract

Railway safety is crucial due to the significant human and economic costs of railway crashes. With more than 1000 accidents reported in the last decade in the United States, railway crashes along the Amtrak remain a major concern. This underscores an ongoing safety challenge that the industry must urgently address. How do network centrality measures explain traffic safety in transportation networks? In this work, we conduct an empirical investigation on the power and significance of the impact of four network centrality measures on the frequency of crashes in the USA's railway network. Leveraging two datasets on railway crashes at U.S. highway-rail grade crossings (HRGC) and Amtrak traffic from 2010 to 2020 and employing advanced spatial econometric models, we show that a station's weighted degree centrality (strength) and local clustering coefficient have no significant impact on the frequency of crashes. Interestingly, there exists a strong and positively significant effect of the weighted betweenness centrality and k-core index of the station on the number of crashes. In this analysis, a crash at an Amtrak station refers to an incident occurring at the nearest highway-rail grade crossing (HRGC) site. Our findings contribute to the growing body of research on transportation safety and offer valuable insights for policymakers and traffic analysts in identifying potential crash hotspots and enhancing safety interventions in the railway network.