User Trajectory And Semantic Analysis Based On Big Data

With the arrival of wireless mobile Internet era, mobile phone that many people carry every day record their activities anytime and anywhere.These massive mobile data contains a wealth of temporal and spatial information, offering a new perspective and limitless possibilities for research of human behavior patterns and urban space environment. At the domestic level, the mobile communication network in the main city of China has achieved 100 percent coverage, mobile phone users will leave electronic "footprint" in the mobile communication operators' server when they use mobile phone in any area of the city. These rich data provides a abundant resource for study of user mobility patterns, grasping the pulse of the city, and analysis of the relationship between user behavior and city space. The huge value of mobile communication data is attracting wide attention from all walks of life.Based on the above background, this paper takes the mobile communication network data as the basis, and studies the relationship between the user's mobile behavior and the characteristics of the urban area, the main contents of the study can be summarized as follows:(1) Application of mobile communication data to some fields, such as population distribution, population movement patterns and urban functional characteristics is reviewed. The related papers are summarized in this paper. The theoretical knowledge and main analysis methods of the research are studied, and the shortcomings and defects of the research are summarized. Finally, the research ideas of this paper are formed.(2) This paper has studied the method of data cleaning with the data set used in this paper, completing the data distincting, denoising and other preprocessing operations with the help of Hadoop platform and the numerical analysis toolkit of Python, such as: pandas, numpy.(3) Modeling the base station traffic with time series, and with the help of Matlab and Eviews,extracting the features of the base station time series from the three aspects: statistical characteristics, flow sequence features in time and frequency domain. And implements k-means clustering algorithm using python's machine learning kit sklearn,realizes semantic calibration on the base station.(4) Combined with the semantic and spatial characteristics of the base station,realize the automatic recognition of urban functional areas by using k-means algorithm, and the recognition efficiency is verified.(5) Mapping the users' trajectory to the urban function area, the OD(Origin-Destination) matrix is constructed and visualized by map analysis tools such as MapInfo, CartoDB and Mapv, then the macro semantic and patterns of the user's trajectory is analyzed.(6) Abstract With the directed graph, the OD network is abstracted,and the close degree of the urban area is evaluated by the connectivity index of the graph.Based on the above theoretical research, this paper analyzes the user's mobile law deeply, put forward an effective model to identify the urban functional area, and give the method of evaluating the functional area. The research results have reference value for city planning, traffic control, road design; at the same time, this work in the city the population mobility can provide theoretical research ideas for the study of human dynamics, can also provide the location of the commercial service scheme based on.