Data Mining to Improve Planning for Pedestrian and Bicyclist Safety (01-003)

Project Description: The study goal was to identify high-risk signalized intersections for walking and bicycling in the City of San Diego using data mining methods. The data used in this study was collected from multiple sources such as San Diego’s automated pedestrian and bicyclist counting system in 2015, several video cameras through National Data and Surveying Services (NDS) in June and July 2018, crash data from SWITERS from 2006 to 2016, and GIS shapefiles. Data Scope: The data from multiple sources as mentioned above were compiled to create a data table in csv format. Total number of observations for this table is 1520 with a total of 502 variables (i.e., columns). Data Specification: A detailed description of each variable in data set can be found in the file titled: SafeD-01-003-Appendix E- DataSpecification.