Spatial-Temporal Analysis And Platform Based On Instant Communication Service

In recent years,technology and information industry are developing at a high speed.Mobile devices represented by mobile phones are rapidly spreading.The mobile Internet business involves all aspects of life.Humans are increasingly dependent on mobile devices such as mobile phones.The traffic data collected by mobile network operators is a rich source of information about human habits.Through analysis and research,we can understand the characteristics of the city,promote urban planning,transportation planning,etc.On the other hand,mining Internet business information and user behavior characteristics can help enterprises explore more potential users.In this paper,we mainly use the instant communication data parsed from the DPI data provided by the mobile operator to build a big data analysis and processing platform,analyze the behavior characteristics of users of instant communication services,and conduct urban network activities from time and space dimensions.The main contents include:First,we have developed a cellular wireless network big data platform.Based on the current mainstream big data processing technologies Hadoop and Spark,and based on the storage and computing requirements of mobile cell data,we have built a big data platform for storing and processing cellular mobile data.We conduct development based on open source big data components.We implemented four tools,Spark Query,Job Submit,Autoflow and Data Transfer.And we divided the big data processing platform into three modules:data storage and transmission module,data processing module and data workflow module.Big data analytics systems feature big data storage,data transfer,multiple ways of data analysis,and automated workflows.Finally,this paper tests the task execution efficiency of the big data processing platform.The results show that the Spark Query processing speed achieved in this paper is about six times faster than the Hive processing speed.Second,the analysis of the distribution characteristics of instant communication services:Based on the instant communication service data,we analyze the distribution characteristics of the user in the time dimension and the base station in the spatial dimension.Firstly,comparing the multiple services in the cellular network data,it is found that the instant communication service has many characteristics such as active users,wide moving range,long active time and scattered usage time.Then,in the time dimension analysis,We analyzed the distribution characteristics of user session duration.The distribution is found to be consistent with the cut-off power law distribution.We analyze the distribution characteristics of the time interval of users using instant messaging services and find that they conform to the power law distribution.We statistically analyzed the characteristics of the number of active users,the number of records,duration and traffic,and found the periodicity of users in time.Finally,in the spatial dimension,we find that the over-lap rate of the mobile user's mobile trajectory in the instant messaging service on two business days is higher than the overlap rate on other dates.We studied the density of base stations and HTTP records in urban and suburban areas.It is found that the Log-normal distribution fitting result is best.And there is a linear relationship between the density of the base station and the number density of the recording strips.Third,Spatial-Temporal analysis method of instant communication service:We use the exploratory factor analysis(EFA)method to analyze the instant communication service in the cellular network.In the time dimension,we analyzed the laws of cellular network activity and found activity characteristics with typical working days and rest days.Each of the working days includes the characteristics of working hours,lunch and dinner hours,and evening breaks.In the analysis of geographic regions,we have found areas with different types of urban functional types,including residential areas,work areas and entertainment areas.And different regions have different patterns of activity in time.By using factor analysis of different statistics such as traffic,number of users,number of HTTP records,and duration,we found the versatility of EFA in the spatial-temporal analysis of cellular network data.We analyze different cities and find that the more developed the city and the more data,the more laws that the method can reveal.