Modeling bicycle volume using crowdsourced data from Strava smartphone application

Abstract

Cycling as a healthier and greener travel mode has been more and more popular among citizens especially for short-distance trips. Since cycling has become an efficient way to reduce energy consumption, eliminate traffic emissions, and improve public health, it is critical to estimate bicycle volume on each roadway segment if possible, and to explore its potential impact on cycling. Therefore, this paper utilizes a prevalent crowdsourcing-based data collection method to model the bicycle volume in the City of Charlotte, North Carolina. The data are aggregated by Strava Metro from the users’ smartphone application. To process the data, essential information regarding manual count bicycle volume, crowdsourced bicycle data, road characteristics data, sociodemographic data, zoning data, temporal data, and bicycle facility data are combined using both the ArcGIS and SAS. After the data processing step, two linear regression models are developed to quantify the relationship between bicycle manual count data and Strava Metro bicycle data as well as other relevant variables. Modeling results are analyzed and bicycle volume on most of the road segments in the City of Charlotte is estimated. A map illustrating the bicycle ridership in the City of Charlotte is also created.

Keywords

• Bicycle volume
• Crowdsourced data
• Manual count bicycle data
• Linear regression models