An Exploration of Bicyclist Comfort Levels Utilizing Crowdsourced Data
Bicycle transportation has become a central priority of urban areas invested in improving sustainability, livability, and public health outcomes. Transportation agencies are striving to increase the comfort of their bicycle networks to improve the experience of existing cyclists and to attract new c...
| Main Author: | |
|---|---|
| Format: | Others |
| Published: |
PDXScholar
2015
|
| Subjects: | |
| Online Access: | https://pdxscholar.library.pdx.edu/open_access_etds/2529 https://pdxscholar.library.pdx.edu/cgi/viewcontent.cgi?article=3534&context=open_access_etds |
| Summary: | Bicycle transportation has become a central priority of urban areas invested in improving sustainability, livability, and public health outcomes. Transportation agencies are striving to increase the comfort of their bicycle networks to improve the experience of existing cyclists and to attract new cyclists. The Oregon Department of Transportation sponsored the development of ORcycle, a smartphone application designed to collect cyclist travel, comfort, and safety information throughout Oregon. The sample resulting from the initial deployment of the application between November 2014 and March 2015 is described and analyzed within this thesis. 616 bicycle trips from 148 unique users were geo-matched to the Portland metropolitan area bicycle and street network, and the self-reported comfort level of these trips was modeled as a function of user supplied survey responses, temporal characteristics, bicycle facility/street typology, traffic volume, traffic speed, topography, and weather. Cumulative logistic regression models were utilized to quantify how these variables were related to route comfort level within separate variable groups, and then the variables were used in a pooled regression model specified by backwards stepwise selection.
The results of these analyses indicated that many of the supplied predictors had significant relationships with route comfort. In particular, bicycle miles traveled on facilities with higher traffic volumes, higher posted speeds, steep grades, and less separation between bicycles and motor vehicles coincided with lower cyclist comfort ratings. User supplied survey responses were also significant, and had a greater overall model variance contribution than objectively measured facility variables. These results align with literature that indicates that built environment variables are important in predicting bicyclist comfort, but user variables may be more important in terms of the variance accounted for. This research outlines unique analysis methods by which future researchers and transportation planners may explore crowdsourced data, and presents the first exploration of bicyclist comfort perception data crowdsourced using a smartphone application. |
|---|