RSF 77, HSIP - Quality Control of Crash Locations

Crash location data has a strong impact on decisions made to improve highway safety. Not only is the crash location important, but the roadway geometry also plays a role in prioritization and identification of high-risk roadway segments and intersections. In Kentucky, the accuracy of crash locations has been found to be very high (about 92% of crashes are located within 500 feet of where they occurred) (Green and Agent, 2011). This is a significant improvement over the 50% accuracy observed years ago, before the addition of a GIS-based map to the crash form (Green and Agent, 2004). While the current location accuracy is high, there is concern regarding the 8% of crashes that are incorrected located. The objective is to investigate crash location errors and develop an algorithm to update the county, route, milepoint, latitude, and longitude to the most likely crash location. Any crash that is identified as being plotted differently by milepoint as compared to latitude and longitude will be flagged. Manual review will be used to identify patterns in these errors. A methodology will be developed to “select” the correct location. This methodology will be automated such that it can be used annually with the KTC crash extract file. The idea here is that all crash analysis conducted by KTC will use the corrected location data.