Building a crash location database should be simple: locate the events on a map. That part is simple, but not as accurate and complete as it could be. To increase the usefulness of California’s statewide crash data a team added pre-processing and cleaning, an offset tool and a new statewide mile marker database to the mix. The outcome? An overall geocoding match rate of 91% and a clean database to serve “clinicians, researchers, policymakers, and practitioners in the fields of traffic safety and public health.” Read up on what a broad-based international team learned in its research.
How could creating a database of crash locations be difficult? In California police reports are written with relative positions - based on mile markers or intersections. In addition, the data had to be pre-processed and cleaned, an offset tool needed to be built, and statewide mile marker database developed. The outcome? An overall match rate of 91% and a clean database to serve "clinicians, researchers, policymakers, and practitioners in the fields of traffic safety and public health." Read up on what a broad-based international team learned in its research in this article published in the International Journal of Health Geographics.
Authors: John M Bigham, Thomas M Rice, Swati Pande, Junhak Lee, Shin Hyoung Park, Nicolas Gutierrez and David R Ragland

