Design And Implementation Of Gathering Patterns Analysis System Based On User Trajectory Data

With the development of information technology,all aspects of people's way of life are deeply influenced by the network.For examples,basic necessities of life,like shopping,getting out,are penetrating by the network.Besides,way of socializing,educational mode,entertainment projects has been great changed.In this context,the system focuses on the spatiotemporal data of users in network logs.Extract the location datas and divide them according to time to model user trajectory,analyze features of gathering mode,and discover the group event of the network users.This system proposes analysis group gathering mode based on user trajectory data.It describes the gathering information from the aspects of starting time,ending time,number of participants,participator counts,gathering status and so on.Based on Map Reduce parallel processing calculation model and Spark large data processing engine,design and implement clustering operator and discovery algorithm.Using the time slice technology and linear interpolation to model the user trajectory,get standard,complete user trajectory data.The DBSCAN clustering algorithm is designed to apply the user location data of a single time slice.The cluster of user location data with basic user information satisfy the density requirement which is obtained.A crowd discovery algorithm for scale,density,duration and geometric stability is designed and implemented in parallel computing environments.The algorithm of gathering discovery is verified on the basis of discovered crowd.On the basis of getting analysis gathering results,the information display platform is designed to play gathering information in the form of Web.Through the access to the plug-in of GIS service provider,achieve friendly interface which is simple and easy to use,Redis is a real-time rendering database because of its fast read and write features to support the real-time presentation of gathering results,MySQL is a historical storage database,to achieve the storage function of historical analysis results.Finally,we design test methods of the system,including functional testing and performance testing.In the testing process,we optimize the program to achieve industrial conditions.The system has been in normal operation.