Design And Implenentation Of Residens' Travel Hot Routes And Attractive Areas Discovery System

In recent years,with the rapid development of GPS devices,wireless communications and satellite technologies,and the widespread use of mobile Internet,more and more trajectory data can be easily obtained.At the same time,these data contain a lot of useful information,which can effectively map people's travel information,activity rules,etc.Therefore,how to use new technology to effectively mine the knowledge of trajectory data has been the direction of further research.The main purpose of this paper is to analyze the distribution of hotspots and areas of users' travels in Beijing.That is,by statistically analyzing the trajectory data of the residents in the city,referring to the density and time factors,the key segments such as longitude,latitude and time information are extracted from the trajectory data,thereby determining the effective staying position of the user in the city,and thereby identifying the user's travel.Routes and attractive areas,from which the changing patterns of people in a specific space-time environment are excavated.At the same time,in the environment where the data scale is increasing,the existing hotspot route and attractive area recognition processing are inefficient.The data is processed by the DBSCAN clustering algorithm based on K-D Tree index structure and Spark GraphX.The system uses the mobile Internet user location trajectory data as the source data,and divides the data cleaning into two steps.First,the null and duplicate values appearing in the data are regarded as regular dirty data,and then for the proprietary ping-pong data in the base station data,the cause is analyzed and the improved sliding window method is used to process it.Since the trajectory data is spatio-temporal data,a time-space dual-dimensional staying point extraction algorithm is designed for this characteristic,and a suitable number of stay points containing user location data and its basic attributes are extracted.The stay point is divided into several sub-track segments as feature points.Based on this,the DBSCAN clustering algorithm based on Spark GraphX is designed and implemented.The cluster segment and the starting point are clustered res pectively to obtain the inclusion of the set threshold.The cluster of user location data and basic attributes obtains the hotspot route and the travel hotspot attractive area of the residents in different time periods.Based on the results obtained,the i ntermediate results and the final results are saved to the database.Finally,the platform is displayed by calling Baidu map interface,and the results are displayed in the form of Web,and the obtained stay points and the user's travel hotspot routes and attractive areas are projected onto the map of Beijing.