Associations between Road Network Connectivity and Pedestrian-Bicyclist Accidents

It has been extensively accepted that the road network connectivity can positively impact the propensity and duration of non-motorized travel. But its impact on non-motorist traffic safety is still under debate: on one side, well-connected road network could lead more through traffic into the core area of a region so that pedestrians and bicyclists would be more frequently exposed to conflicts with cars; on the other side, it could be safer when vehicle speed is slowed down by dense intersections and drivers are forced to concentrate on surroundings by active walking and bicycling. This debate stimulates the paper to estimate the associations between road network connectivity and pedestrian-bicyclist crashes. Four commonly utilized connectivity measures including block density, intersection density, street density, and mean block length are calculated based on the road networks of 321 census tracts in Alameda County, California. Then the four measures together with other factors like traffic behavior, land use, transportation facility, and demographic feature are employed separately in a spatial statistical model called geographically weighted regression. Conclusions are: first, the decrease of pedestrian-bicyclist accidents is significantly related to higher block density, higher intersection density, higher street density, and shorter mean block length; second, compared with the other three connectivity measures, street density is better for modeling because of its higher stability and stronger explanatory ability; third, employing street network, traffic behavior, and transportation facility data into the same model can produce the best model fitness.