一种面向时序数据的交互式可视化系统

An Interactive Visual System for Time Series Data

在科学、经济、工程、工程等各领域,每时每刻都产生大量时序数据。挖掘多维时序数据中蕴含的相关性特征和周期性规律,是可视分析领域研究的重点。本文在总结了过往案例和可视化方法的基础上,提出了一种称为时序螺旋的可视化交互系统。该系统基于时间粒度和时间原语的概念进行设计,旨在通过可视分析的方法,挖掘数据维度之间的相关性关系和周期性规律。时序螺旋综合了年轮图、族线等多种周期性数据可视化布局方法,并辅以多种交互模型互为补充。最后以实际数据集的分析案例展示了我们的方法在探索和理解多维时序数据的有效性。

Massive data has been produced all the times,in the fields of science,economy,engineering and so on. Mining the correlation characteristic and periodic law contained in the multivariate time series data,is a key topic in the visualization community. On the basis of summarizing the past cases and visualization methods,this paper presents a visual interactive system called Time Spiral. The system is designed based on the concept of time granularity and time primitives,aiming to excavate the correlation between the data dimension and the periodic law through the visual analysis method. Time Spiral integrated the annual growth ring,cluster and other periodic data visualization layout methods,supplemented by some interaction models. Finally,we use two case analysis which are based on real data sets to demonstrate the effectiveness of our approach in exploring and understanding multidimensional time series data.