Analysis of Bike-Sharing Travel Behavior Using Multi-Day O-D Trip Data

Harvesting O-D trip data from Tianjin’s bike-sharing system, this paper analyzes the travel behavior of bike-sharing in Tianjin from a regional perspective, which offers significant insights into understanding the spatiotemporal flow characteristics of bike-sharing demand. First, the spatiotemporal travel behavior of bike-sharing is analyzed from a regional perspective. Second, this study employs the non-negative matrix factorization (NMF) algorithm to classify bike-sharing travel patterns into two dimensions, namely spatial and temporal. Finally, the Poisson regression model is used to interpret each pattern. The study presents a series of findings, such as the spatial and temporal distribution patterns of bike-sharing travel behavior and five different interpretations of bike-sharing travel patterns in Tianjin. The findings contribute to understanding bike-sharing behavior, predicting bike-sharing system demand, managing its services, and promoting green mobility in cities.