Analysis Of Passenger Flow Behavior Based On Big Data Related To Shenzhen-hong Kong Ports

With the continuous development of reform and opening up,the exchanges between Shenzhen and Hongkong are deepening in many aspects of culture and economy.In the process of continuous deepening of communication,there are many potential problems,such as traffic congestion,smuggling of water tourists and the decline of service quality of tourists.Based on big data platform of Hadoop and data processing technology,this paper sets up a unified space-time standard database,analyzes port passenger traffic in different periods from time axis,analyses the movement track of different people from the space axis,and quantificationally analyses the life of different time and different crowd in port area through spatial-temporal data mining technology.In order to optimize the urban intelligent traffic management,improve the efficiency of the port operation and improve the traveler’s travel experience,the dynamic characteristics and laws are used to explore the travel behavior patterns of different people.The main research work of this paper is as follows:(1)Data preprocessing technology is used to complete data fusion of public transport data,and combined with time and space mining technology,the crowd segmentation and mode analysis of port passenger flow are carried out.On the basis of the big data platform,the GPS data drift points are removed.Based on the travel chain estimation algorithm,the station points are calculated under the conventional bus traffic.The spatial and temporal characteristics of the passenger flow are analyzed by the preprocessed data,and the corresponding characteristics are extracted and analyzed by the hierarchical clustering algorithm for the population division and model analysis of the port passenger flow.In order to solve the problem of slow operation efficiency of hierarchical clustering in large amount of data,a solution based on density clustering and hierarchical clustering is proposed,and multiple similarity is combined to improve the efficiency of the algorithm.(2)Using the captured Hongkong mobile station data and the POI information point(Point of Interest,that is,"interest point")to establish a label for the region,and use the communication data to analyze the tourist’s temporal and spatial trajectories.Using the map label classification method,the Hongkong POI data captured by the Google map are divided into 16 categories,and the base station data and related formulas are used to label the location of the base station,and the spatial and temporal trajectories of the population are analyzed from the different dimensions using the data of the single cell phone card and the base station information fusion.(3)Clustering algorithm is used to segment tourists’ destination preferences.Combined with 16 categories of classified labels,the feature vectors of travel time consumption are established.After weighting and dimensionality reduction,the Kmeans clustering algorithm is used to divide the tourist destination preference.In order to avoid the local optimal solution,an improved algorithm based on the density optimization is proposed to select the initial cluster center.The local density of the data points and the distance from the point to the point with higher local density are selected,and the point at high density is chosen as the initial clustering center.Then,according to the division of business circle,we use association rule algorithm to find out the relationship between major business circles.This article analyzes the characteristics and laws of the activities of the population in the deep harbor by multi-source data,and it has certain guiding significance for optimizing the urban traffic allocation and improving the service quality of tourists.