Mining Spatiotemporal Characteristics of Car-sharing Demand

Since car-sharing demand plays an essential role on the management and service of car-sharing system, this paper attempts to analyze the temporal and spatial characteristics of the car-sharing demand, aiming to discover spatial and temporal patterns and association rules. Based on a clustering algorithm (i.e., DBSCAN), the spatiotemporal characteristics of car-sharing demand are studied. On the basis, the demand is divided into various clusters in space. Furthermore, the correlations between any two clusters are studied. The results show that car-sharing demand is high on Friday, Saturday and Sunday, and has strong correlation with time and space. It is expected the results can support the management of car-sharing system, and further promote the service level for passengers.