K-means clustering method based on kernel density estimation to analysis residents travel features: A case study of Chengdu

Abstract

In order to study the spatiotemporal features of urban centre residents during peak hours during workdays and rest days, based on taxi on and off location information and urban points of interest data, a Geographic Information System (GIS) Kernel density estimation (KED) is used. Combined with the K-means clustering algorithm, the peak hours of residents 'travel and hotspot areas for boarding and alighting are identified, and the strength of the interaction between residents in each area is analysis using the structured Georgy Voronoi, and the spatiotemporal features of residents' travel are summarized.