Characteristics Exploration of Hospitals in Shanghai through Spatial Relationship Using Taxi Data

With the increasing volume of spatiotemporal traffic data, many scholars used it to explore the characteristics of urban facilities. In this paper, a method of classifying the main tertiary hospitals in Shanghai using taxi data of online car-hailing was proposed. Rasterization of the researched area and extraction of OD (Origin-Destination) records were done, and an optimum buffer distance threshold was determined. By comparing the previous studies, we designed the clustering approach (based on K-Means algorithm) and set new inputs based on trips at different distances from a researched hospital. A reasonable result was achieved: 4 types of hospitals are obtained and a discussion was conducted. This research is valuable for identifying the function and orientation of hospitals, and can help government better understand the spatial relationship of hospitals.