Transportation Safety Planning: A Spatial Analysis Approach

Helai Huang,Pengpeng Xu,Mohamed Abdel‐Aty

2013-01-01

Abstract:In the past decade, considerable efforts have been made to explore various safety conscious planning approaches to form the next-generation planning method, i.e. Transportation Safety Planning (TSP). This study proposes a spatial analysis approach for TSP, with four phases, i.e. Data preparation, Examination of Modifiable Areal Unit Problem (MAUP), Regionalization for Traffic Safety Analysis Zone (TSAZ), and Development of TSP-Prediction Model (TSP-PM). To illustrate the approach, the authors present a preliminary analysis using Hillsborough County data of Florida (2005-2007). Based on the existing TAZ delineations, they re-regionalized the whole county with zone numbers at an incremental step-size of 50, resulting in 14 zoning schemes. In the regionalization, a multivariate combination of key factors was used as the regionalization criteria, i.e. severe crashes, total road length and total trip production with weights 0.5, 0.25 and 0.25, which respectively reflect the level of safety, transportation facilities, and trips as a major concern at the planning level. Accompanied with a univariate MAUP examination, a total of 64 exploratory TSP-PMs were developed based on the 14 new zoning schemes and three major categories of factors, i.e. intersection density, road length with different speed limit, and trip generation factors. Results showed that as the number of zones increases, the safety interpretation power by transportation system factors decreases while the spatial autocorrelation increases. To the contrary, better model-fitting performance is associated with larger number of zones. Therefore, rational selection of zoning schemes would have to depend on detailed requirements for specific usage of zoning partitions. This research effort is expected to achieve a substantial step toward a new generation of spatial decision support system for TSP practice.