基于组件协同的时空数据可视分析系统

A visual analysis system for spatiotemporal data based on components coordination

设计实现基于组件的、多视图协同的可视分析系统,用于探索分析复杂的时空数据,挖掘其中隐藏的空间模式、时间模式及多维属性模式.系统以杭州房地产数据为例进行分析研究,该数据是一种典型的时空数据,包括楼盘的空间分布情况、均价和销量变化规律、楼盘属性相关性等复杂特性.系统以可视化的方式,提供了综合有效的、可交互的分析方法.系统开发实现4个核心可视分析组件:地图组件、平行坐标组件、时间序列组件和treemap组件.地图组件提供空间视图;平行坐标组件提供多维视图;时间序列组件提供时间视图;treemap组件提供综合视图.系统以一种可扩展、可定制的方式,实现各个组件之间高效的关联协同,从而大大提高了系统整体综合分析能力.

In the paper, a component-based visual analysis system with multiple coordinaed views is presented. It is designed for exploring and analyzing complex spatiotemporal data, and mining hidden spatiotemporal and multivariate attributes patterns. Here, Hangzhou real estate dataset is taken as example data for analysis. This dataset is a typical spatiotemporal data, with characteristics of the spatial distribution of house, average price and sales variation pattern, the correlation of house attributes, etc.. This system provides an effective and interactive analysis method in visualization way. The four core visual analysis components are developed： GeoMap, Parallel Coordinate Plot （PCP）, Time Series and Treemap. The GeoMap provides spatial view. The PCP provides multi-dimensional view. The TimeSeries provides temporal view and Treemap provides comprehensive view. The whole system is designed in a extensible, customizable way and various components are coordinated efficiently among each other. Thus the overall system comprehensive analysis capabilities will be improved greatly.