Safety impacts of the discrepancies and accesses between adjacent traffic analysis zones

The analyses of the safety effects of the road network features can help to understand the risk factors in road crashes and provide useful policy implications. In macro-level analyses, adjacent traffic analysis zones (TAZs) could influence each other on road safety. Therefore, it is necessary to move beyond the common practice of just examining the road network of the TAZ itself and analyze how and to what extent the interrelated factors of the objective TAZ and its adjacent TAZs interact to affect crash frequency. This study aims to investigate the safety impacts of the discrepancies and accesses between adjacent TAZs. Based on data from London at the level of Lower Layer Super Output Area (LSOA), the Bayesian model was employed to analyze the safety effects of various factors related to the discrepancies and accesses (i.e., access number, discrepancies of road network density). A spatial weight based on the access number was applied when establishing the Bayesian spatial models. Results suggest that TAZs with greater discrepancies and more accesses from adjacent TAZs experience more crashes.