Research On The Model And Visualization Method Of Trajectory Big Data Based On Hadoop

Large trajectory data contains rich space-time information,is a sampling sequence with location and time information,and contains the space-time dynamics of the collected object.Compared with the traditional 4V features of large data,large trajectory data has more features: Veracity(true-false coexistence),data missing,noise points,redundancy and other anomalies commonly exist in large trajectory data;Visualization,large trajectory data often need to be combined with the actual spatial location and spatial information of data to make full use of map pairs of operation status,urban road network characteristics,and so on.Visualize the presentation.In recent years,with the continuous development of satellite positioning technology,LBS technology and the Internet,location data has been collected in a variety of ways,large trajectory data is explosively growing.Large volume of trajectory data is generated at any time of the day,which leads to the general database can not better adapt to large trajectory data with real-time expansion capacity,and large trajectory data has multiple sources and numbers.Due to the complex structure and other characteristics,researching a storage and management technology suitable for large trajectory data has become the focus and difficulty of current research.Visualization is a way of data expression,which is often used in data analysis and data mining.Due to the large volume of data in large trajectories,visualization often encounters such problems as long visualization time,Page Carton,high data density in hot traffic areas,poor visualization of heatmap,hot core phenomenon and general visualization effect.Based on the large data of Beijing multi-source transportation,this paper builds a Hadoop distributed platform to calculate and store large data of transportation,improves the visualization of large data of trajectory by clustering large data of trajectory and adding Geohash codes,and establishes a visualization system based on Web GIS for data visualization testing.This paper mainly studies the following aspects:(1)Build Hadoop Distributed Storage Framework.Build a micro-distributed cluster with lab computers,pre-process the data,realize taxi track mining and OD information statistics in different regions at different time periods,and visualize the results with front-end web open source map library.(2)Construct taxi track data model.Based on HBase,a track data storage model is established.For the visualization of taxi track data,a clustering method is proposed to process large data of original taxi track,which reduces the amount of visualization drawing calculation and data transmission.Geohash encoding is added to the original taxi track data to improve the query efficiency of track data in a small area.Experiments on taxi track data sliced at a certain time show that the loading efficiency of thermogram increases by about30% when the zoom level is higher,and the query speed of scatter query based on Geohash coding increases significantly when the zoom level is higher,which verifies the feasibility of this scheme.(3)Build the storage model of electronic map.By setting up the HBase distributed array storage data warehouse,combining the storage characteristics of electronic map data and HBase column family,an electronic map storage model for service web visualization is built,which improves the loading efficiency of map raster data.After 100 query comparisons of map data,it is found that the query efficiency increases by about 20%,and has a stable real-time loading rate.(4)Set up visualization platform of traffic big data.Based on the above research and data base,the system framework and the query and visualization process of each data of the large traffic data visualization platform are designed.Through Java background code to achieve the operation and data transfer of Hadoop framework,combined with the Java SSM framework(Spring+Spring MVC+My Batis)to complete the data transfer and access of web pages on the service side and front end,the front end uses the open source Java Script map engine to display the spatial location information of large traffic data,combined with echarts,d3.js and other chart libraries to visualize the basic information of the data.