Construction of spatial-temporal portrait database of passenger flow in urban rail transit station based on data drive

In order to relieve the urban traffic pressure and meet the needs of refined passenger transportation organization, this paper describes and summarizes the temporal and spatial characteristics of passenger flow at existing rail transit stations, constructs a set of unique station temporal and spatial portrait library, realizes the fine division of stations, and provides a basis for the passenger transportation organization at stations. Firstly, the passenger flow data of automatic ticketing system (AFC) is preprocessed, and the passenger flow characteristics are described and classified by K-Means ++ based double-layer clustering method combining the passenger flow scale and the passenger flow time series characteristics. Secondly, based on the source-sink theory of landscape ecology, a normalized location-weighted landscape index (NLWLI) is established to evaluate the occupational and residential land around the classified rail transit stations, and then a complete set of spatial and temporal portrait database of station passenger flow is constructed to realize scientific and reasonable station classification. It is of great significance for passenger flow forecasting, station facility layout optimization and surrounding land development and construction.