多维时空数据协同可视分析方法

Collaborative Visual Analytics of Multi-dimensional and Spatio-temporal Data

随着获取渠道的多样化以及采集过程的规范化,数据普遍具有多维、时空属性.由于不同的属性维度中数据的分布方式和表达意义各不相同,针对数据的多维时空属性开展综合的分析和全面的探索存在较大的困难.为此,提出一种多维时空数据协同可视化方法.首先设计属性关联视图展现空间对象的多维属性数据及其关系;然后利用MDS算法将原始多维属性数据按照时间顺序关系分别降解至一维空间构建时序平行坐标系,并且对各时间轴的投影坐标点进行布局优化和层次聚类操作,利用信息熵度量类别的稳定性对平行坐标系中点和线的颜色进行映射;最后设计圆状布局的层次链接图,展示感兴趣特征的层次结构关系.文中通过设计便捷的交互模式有效地关联各个视图的可视化设计,构建多维时空数据协同可视分析系统.GDP数据和空气质量监测数据的实例分析结果表明,文中方法能够综合考虑多维时序属性,帮助用户快速探索数据中隐含的特征模式;专家用户的反馈进一步验证了该方法的有效性和实用性.

With the diversification of data acquisition and the standardization of collection process,the obtained datasets always contain multi-dimensional and spatio-temporal attributes.Because of the various distributions and expressions of the attributes,it is a difficult and tedious task to carry out an integrated and comprehensive analysis for multi-dimensional and spatio-temporal data.Therefore,we propose a collaborative visualization approach of the multi-dimensional and spatio-temporal dataset in this paper.The attribute association view is designed to present the dimensional properties and their relationships of a spatial object of interest.By means of multi-dimensional scaling,the original dataset is degraded to one-dimensional space in a chronological order,which further enable the construction of the sequence parallel coordinate system.In order to enhance the visual perception of the parallel coordinate system,the coordinates can be overturned to avoid vigorous quivers of the lines and further clustered to achieve different features,the stability of which is measured by entropy and used to define a meaningful color mapping.A circular node-linked diagram is designed to present the hierarchical relationship of the features of interest.Based on the convenient interaction modes,the proposed approach can effectively correlate the above visual designs and construct collaborative visual analysis system for the exploration of multi-dimensional and spatio-temporal datasets.It can be concluded that the proposed system is able to help users quickly explore the characteristic patterns hidden in the multi-dimensional and spatio-temporal dataset through the actual case studies of GDP data and air quality monitoring data.The validity and practicability can be further demonstrated based on the feedback of the experts.