Gravity model in dockless bike-sharing systems within cities.

Due to previous technical challenges with the collection of data on riding behaviors, there have only been a few studies focusing on patterns and regularities of biking traffic, which are crucial to understand to help achieve a greener and more sustainable future urban development. Recently, with the booming of the sharing economy, and the development of the Internet of Things (IoT) and mobile payment technology, dockless bike-sharing systems that record information for every trip provide us with a unique opportunity to study the patterns of biking traffic within cities. We first reveal a spatial scaling relation between the cumulative volume of riding activities and the corresponding distance to the city center, and a power law distribution on the volume of biking flows between fine-grained locations in both Beijing and Shanghai. We validate the effectiveness of the general gravity model on predicting biking traffic at fine spatial resolutions, where population-related parameters are less than unity, indicating that smaller populations are relatively more important per capita in generating biking traffic. We then further study the impacts of spatial scale on the gravity model and reveal that the distance-related parameter grows in a similar way as population-related parameters when the spatial scale of the locations increases. In addition, the flow patterns of some special locations (sources and sinks) that cannot be fully explained by the gravity model are studied.