Research And Application Of Stay Point Trajectory Data Index Based On Spatio-Temporal Cells

Globally,the data in every trade and aspect of people's lives has become a vital productive force.80% of the data is related to geographical location.Among them,the data used for research on mobile phone's location,data micro-blogging check-in data,etc.,whose location change frequency is low and involving a wide range of objects has become the basis for urban scientific planning and construction.Efficient organization,storage,management,and queries for the large amounts of resident data are among the hot issues that need to be solved urgently.However,the existing spacetime data indexing methods are not suitable for the point trajectory data based on their index structure and application requirements.This paper proposes a scalable index method based on space-time cells for historical data that divides time and space into small compartments of a certain granularity as carriers for storing moving objects.Unlike traditional data-driven indexing methods,this paper's indexing method naturally supports efficient access to a large number of objects due to its spatially driven nature.The index module contains the space-time snapshot index module and the object track index module that can quickly retrieve the distribution of moving objects,and realize space-time range queries,specifically for the moving object track queries and the starting destination queries.To verify the efficiency of the method,the efficiency test experiment was conducted based on the two index modules.The performance and efficiency indicators included the time of building the index,the size of index files and the timeconsuming of common queries.The space-time snapshot index submodule and the object track index submodule were compared with the commonly used 3DR tree and the B+ tree,respectively.Relatively with the 3DR tree index method,the space-time snapshot index submodule has great advantages,especially in the efficiency of building index and querying data.Compared with the B+ tree,the object track index submodule occupies less space,and the trajectory query efficiency has great advantages,the speed of which has been improved about 7 times.Finally,the mobile phone data in Chongqing was used to realize quick retrievals of the personal trajectory and visualization of the traffics at a given space-time condition and the urban residents' analysis of jobs-housing relationships that verified the validity of the index method developed in this paper.