Visual Analysis And Research Based On Data Of Mobile Phone

With the rapid development of mobile communication technology and the network information technology,the communication patterns such as telephone voice and text,video communication,web-page text and voice have rapidly emerged and have produced a large number of multi-dimensional and heterogeneous data.This kind of data reflects people's activity,consumption characteristics from different aspects,but the massive and heterogeneous nature of the data brings great difficulties to the study of human activities.In view of these difficulties,this article adopts the method of information visualization and discusses the visual analysis of user activities based on mobile communication data from the following three aspects:(1)The visual analysis of user behavior based on mobile call data includes two aspects: group-oriented and individual-oriented: for the group-based per capita consumption distribution mining,we use the Pareto principle to rough out all the users,and then use the clustering algorithm K-means to segment the users.Then use the geographical location information to extract the urban residents' consumption levels and the characteristics of group activity areas by heat-map,ect.;A multi-granularity time-series thermodynamic map based on principal component analysis was used to analyze the regularity of community activities and possible event mining.And for individual trajectories,dynamic trajectory analysis is completed.And this paper proposes an individual activity model based on geographic location to build and optimize;For individual social relationships,a probabilistic social network is constructed based on the social tightness and strength model.Aiming at the problem of uncertainty in the range of probability existing in the above network graphs,an improved method based on custom probability nodes was proposed.Finally,the user activity analysis and social relationship analysis were performed through the map service and verified.The method can effectively identify user social relationships.(2)In the visual analysis design of mobile phone user behavior,a time-series heat-map is designed for space attributes and spatial attributes by means of spatial expansion and color mapping.A spatial-temporal contrast visual analysis view based on location information enhancement is designed for centralized or decentralized of the individual activity track.A visual analysis based on multi-view collaborative interaction is designed for the diversity of analysis tasks.A multi-granularity data filtering method based on space-time dimension is designed aiming at the data characteristics of complex and large-scale data dimension.(3)A visual analysis system of mobile phone user behavior based on Django is established.Visual analysis sub-modules system for group and individual mobile user behavior were developed as the diversity of objects.Taking the records of 20562730 for 5 months as an example,this paper exploits the temporal movement pattern of group mobile phone users,the distribution of urban per capita consumption,the level of urban consumption,the regional characteristics of consumer groups activities,etc.;and it excavates the activity trajectory,activity model and optimization,activity range,activity location,activity area probability of individual call user behavior As well as individual social networking,etc.Which verify the effectiveness of the system.