Optimization And Implementation Of Service Technologies Of Spatio-Temporal Big Data Applications For Tourism Scenarios

Spatio-temporal big data is big data with two dimensional attributes of time and space.In vast amounts of data,more than 80% of the data is related to time and spatial location.In the application field of space-time big data,the tourism industry has a very broad prospect.By using related technologies,travel agency teams can carry out tracking monitoring and scenic load monitoring.Also,scenic area managers can can monitor the real-time load of the scenic spot through the passenger flow heat map.What's more,they can also timely gather the on-site data of individual visitors that have been booked in advance to the park on the Internet,that tourists can judge whether it is suitable to enter the scenic spot within a period of time according to the statistics of the number of people and the carrying capacity of the scenic spot.Based on the tourism intelligence monitoring platform,this thesis studies the optimization of some key technologies in spatio-temporal big data,which mainly involves three aspects: spatio-temporal acquisition performance optimization,distributed storage design and the nearest neighbor query optimization based on moving objects.First,this thesis designed a reading and writing framework which called STClient based on the big data platform to improve the collection performance of tourist flow data.we introduced its overall process,and then,according to the characteristics of spatialtemporal data,we optimized it from five aspects of file reading and writing,data preprocessing,spatial-temporal index scheme design,serialization and warehousing.Then,this thesis designs a distributed storage system for spatio-temporal data.We designed a project-based system architecture according to the spatial and temporal characteristics of spatio-temporal data.And then,by dividing the date by temporal and spatial attributes,we build a spatial heat tree to improve the performance of system load balancing.In addition,on the basis of the existing cold and hot data separation scheme,this thesis makes a further improvement to improve the availability of the storage system and the performance of the system.These schemes enable the system to effectively support the subsequent series of passenger flow data analysis.Finally,based on the grid index and the optimized quadtree index,we put forward an index method which can support the nearest neighbor query.By indexing the region where the object is located rather than indexing the moving object itself,the method can exactly reduce the index update times.And then,we proposed a nearest neighbor query algorithm based on the index method to improve the efficiency of the query.