A Cluster Analysis of Uber Request Data via the Transit App in New York City

As ridehailing services like Uber become increasingly common in urban transportation systems, it is necessary to understand their usage patterns. Since private ridehailing companies do not publicly disclose their ridership data, usage patterns can be analyzed using other data sources, such as the Transit app’s Uber request data. The objectives of this research are three-fold: (1) explore the temporal characteristics of Uber requests through data visualization, (2) identify groups of users through cluster analysis, and (3) compare Transit app Uber request data with overall Uber usage data and transit data. The exploratory analysis results suggest that requests occurred most frequently during AM and PM peak periods. K-means clustering identified eight groups of Uber users: long duration and frequent users, off-peak users, PM peak users, AM peak users, party goers, long duration and infrequent users, holiday users, and weekend users. The main trip purposes determined by the clustering analysis were going to social events and to and from the workplace or home. Comparing the Transit app data to the overall ridehailing usage data and transit data suggest that the time distribution pattern of Transit app Uber requests is a combination of transit and Uber usage while the time usage features are more similar to those of Uber users. These results will help transportation departments to better coordinate ridehailing services and public transportation to meet users’ travel needs.

Language

  • English

Media Info

  • Media Type: Digital/other
  • Features: Figures; References; Tables;
  • Pagination: 20p

Subject/Index Terms

Filing Info

  • Accession Number: 01763547
  • Record Type: Publication
  • Report/Paper Numbers: TRBAM-21-03088
  • Files: TRIS, TRB, ATRI
  • Created Date: Dec 23 2020 11:04AM