A Visual Analysis Sistem With Large Scale Taxi Oring Destination Data

With the development of city and the increasing of city population,the modern metropolis city facing serious problems, such as traffic congestion and city resources planning problem. In recent years, with the rapid development of technology and many applications of the sensor technology, a large amount of data has been collected and recorded in all aspects of the state of the city. To solve city problem and to create smart city which based on the analysis of big data has brought new opportunities. This thesis studies the taxi GPS data, and aims to enable the transportation management departments to predict changes of taxi demand in different regions and schedule the taxi departure in advance.This study takes advantages of taxi origin and destination positions for analysis. The system overview includes heatmap and clustering map. The overview can guide users to perform visual analysis in most regions where has some interesting pattern. Combined with the circle pixelgraph, spatial-temporal stacked graph, nested pixel bar, users can discover the knowledge about the differences among taxi flows in each region through interactive exploration. With the help of visual component, analyst can also understand the influences of certain urban planning events exerted on people’s travel pattern from the taxi data.To implement the visual analysis system, we employ a popular platform Hadoop, which is widely used in storing and processing big data. With Hadoop, we can manage and preprocessing taxi OD data efficiently. With B/S architecture, each visual component serves as the front-end data analysis part to analyze the taxi OD data visually. We employ the grid index of taxi OD data to accelerate the query operation when users are interacting with the visual components. Our visual analysis system could facilitate the transportation management departments to discover taxi accumulation pattern and manage vehicle scheduling.