面向GDP数据时空多维属性的可视分析方法

摘要

国内生产总值是经济统计的核心指标,文中提出面向国内生产总值数据时空多维属性的可视分析方法.利用多维标度法对不同时空维度多维属性进行降维,支持用户交互选择时空维度和参与相似度计算的多维属性;利用环状饼图展示产业结构的比例关系,设计垂直指针定位优势产业并动态表达产业结构的变化趋势;利用圆状堆叠图呈现各个产业产值及占比的时序变化信息,堆叠图的绘制顺序进一步提示产业结构的更替特征.设计内圆、中环、外环的可视化模式,实现面向国内生产总值数据时空多维属性的可视分析系统,利用方便而快捷的交互手段关联上述可视化方法,为用户提供快速探索经济统计数据中产业结构时空特征模式的工具.最后通过大量的可视化效果及用户体验结果,进一步验证了文中方法的有效性和实用性.%The gross domestic product (GDP), a significant part of economic statistics, is collected to analyze the development of national economic. In this paper, we propose a visual analytics method to visualize GDP with time-space multidimensional features. MDS is employed to reduce the dimensionality of original multi-dimen-sional data, and the users are allowed to select different perspectives and dimensions to calculate the similarity matrix. We use the circular pie chart to present the relationship of various industries, especially with a vertical pointer designed to locate the major industry. The angle between the vertical pointer and the center line further indicates the development trend of industrial structures. We also take advantage of a centric stream map to reveal the proportion of different industries, and shows the change of the major industry using the drawing order. The user interface of the proposed system is composed of three parts including the inner, middle and outer areas. Sev-eral intuitive interactions are also provided to help users quickly explore the time-space features of the industrial structure in the economic statistic data. A large number of experimental results as well as credible user studies further demonstrate the effectiveness and practicability of the proposed system.