Civil and Environmental Engineering
First Advisor
Jason Hawkins
Date of this Version
12-2023
Citation
A thesis presented to the faculty of the Graduate College at the University of Nebraska in partial fulfillment of requirements for the degree of Master of Science
Major: Civil Engineering
Under the supervision of Professor Jason Hawkins
Lincoln, Nebraska, December 2023
Abstract
Traditionally, traffic prediction has been an engineering challenge due to insufficient quantities of data. In this thesis, the use of a new type of data called location-based service (LBS) data from StreetLight (StL) and a large database of multi-modal traffic counts are used to infer vehicle, pedestrian, and bicycle volumes on all road segments in Lincoln, Nebraska. Additional land use features are incorporated into the model from the EPA Smart Location Database (EPA-SLD). The use of random forest models provides strong measures of fit for all modes, with R2 scores of 0.98 for both the bicycle and pedestrian modes and 0.95 for the vehicle mode. Recognizing that traffic count locations comprise a biased sample, the StreetLight traffic volume indices and EPA-SLD features between our training and prediction samples are compared to one another.
This thesis uses two crash rate measures, a multiplicative rate measure and a ratio rate measure, designed to assess road safety. Using motorized-to-non-motorized crash data from the city of Lincoln and vehicle and active miles predicted from the previously described random forest models, correlation analysis, spatial model analysis and “Relative Rank” analysis are performed. Each type of analysis is designed to compare both crash rate measures to higher level land use data. The results of the spatial model demonstrate varied effects for certain density, diversity and demographic variables. Multiplicative rate correlated strongly with the percentage of low wage workers, employment entropy and multi-modal network density, while ratio rate correlated strongly with multi-modal intersection density and regional diversity.
Advisor: Jason Hawkins
Comments
Copyright 2023, Nicholas Aldridge