Using kernel density estimation in GIS to identify accident hot spots in Melbourne tram stops

Melbourne has the biggest operating tram network in the world with 250 kilometres of double track. There are 1,772 tram stops across the Melbourne tram network and 61% of these stops are kerbside. In addition, around 80% of Melbourne's tram network shares road space with other vehicles, and this causes to exposure risk of accidents for pedestrians at kerbside tram stops. According to Victoria road crash statistics, between 2008 and 2014, 209 pedestrian accidents are occurred when they were walking to/from tram stops or boarding the tram. The aim of this research is to use Geographical Information Systems (GIS) and Kernel Density Estimation (KDE) to find the spatial patterns of pedestrian accidents at Melbourne tram stops. A novel approach is applied in the KDE analysis to identify optimum bandwidth for crash black spot analysis. In addition, by combining KDE and crash frequency analysis in GIS, it is possible to identify pedestrian crash hot spots on tram stops.