Data Requirements for Safety Screening
Dr. Aditya Medury
Prof. Venky Shankar
Dr. Offer Grembek
Dr. Bor-Wen Tsai
California Department of Transportation (Caltrans)
Publications and Resources:
- Identify the Data Requirements for Safety Screening to Identify High Collision Concentration Locations, Technical Report, CA17-2017
California Department of Transportation (Caltrans) intends to implement statistical methods that follow the methodology described in the Highway Safety Manual to identify high collision concentration locations (HCCLs) along the state highway system. A successful implementation of such HCCL identification methodologies, which are referred to as network screening techniques, necessitates the development of safety performance functions (SPFs). SPFs are mathematical equations that relate collision frequencies to traffic volumes at a given location and may include other site characteristics such as road geometry and intersection design. The outcome of an SPF is the expected (i.e., average) number of collisions per year for a given location, and it acts as a baseline to detect whether a site has a “higher-than-expected” number of collisions.
Figure 1. Illustration of different API-based speed limit evaluation approaches
Parallel research efforts to develop California-specific SPFs in Caltrans have used data from the Traffic Accident Surveillance and Analysis System (TASAS), which suffers from some limitations. The first is the absence of data with regards to some attributes (e.g., horizontal and vertical alignment, posted speed limits); and the second is the inconsistent quality of the available data. The extent of this knowledge gap is also non-uniform across different components of the highway system. Hence, the primary goal of this project was to assess these knowledge gaps for SPF development and supplement this information with a thorough review of additional data sources both within and outside of Caltrans. The objective was to develop a roadmap for a data collection plan to facilitate better SPF model estimation, which in turn facilitates better network screening. To devise this roadmap, this report describes the steps undertaken to identify the data needs, evaluate the different data sources that can potentially meet those data needs, and assess their suitability for SPF modeling.
Figure 2. Visual comparison of curves
The outcome of this project was the identification of data sources both within and outside of Caltrans that can be utilized for collecting new variables for SPF development in addition to the data available within TASAS. A suitability analysis framework was also proposed which evaluates the quality of data through the metrics of completeness, frequency of updates and spatial variation. Finally, a roadmap for populating all variables suitable for SPF modeling, either through existing or newly identified sources, was proposed. The recommendations included key performance measures to assess the quality of future data collection efforts, as well as policy considerations to ensure that the data are consistently updated for the entire state highway system.