当前位置: X-MOL 学术J. Comput. Lang. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Visualization approaches for understanding uncertainty in flow diagrams
Journal of Computer Languages ( IF 2.2 ) Pub Date : 2019-04-23 , DOI: 10.1016/j.cola.2019.03.002
Zana Vosough , Dietrich Kammer , Mandy Keck , Rainer Groh

Business Intelligence applications often handle data sets that contain uncertain values. In this contribution, we focus on product costing, which deals with the average costs of product components – that vary significantly based on many factors such as inflation, exchange rates, and commodity prices. After experts estimate the uncertainty information for single items, decision makers need to quickly ascertain the cost uncertainties within the hierarchical data structure of the complete product.

We propose that only a holistic visualization containing both data and uncertainty can provide this kind of quick overview. Such a visualization must be able to visualize tree data structures associated with value attributes. After conducting interviews with product costing experts, we focused on Flow diagrams, which fulfill this basic requirement. However, they need to be extended in order to directly incorporate uncertainty information.

We investigated three visualization techniques applicable to the ribbons of Flow diagrams to convey uncertainty information: Color-code, Gradient, and Margin. Moreover, we designed five visual approaches to show the uncertainty on nodes of Flow diagrams that we evaluated with visualization experts. The approaches add different geometries to the nodes such as triangles, blocks, or forks. The preferred solutions for the nodes was adding forks or filled blocks. With regards to the ribbons, we contribute a user study involving the solution of different product costing tasks using the three different visualizations. Although Gradient was considered an intuitive choice to show uncertainty, it yielded the highest error rates. In contrast, Color-code and Margin were superior depending on the performed task. Based on these findings and the subjective feedback, we designed an integrated approach that combines elements from all three distinct techniques and applied it to Sankey diagrams and Parallel sets.



中文翻译:

用于理解流程图不确定性的可视化方法

商业智能应用程序通常会处理包含不确定值的数据集。在此贡献中,我们将重点放在产品成本计算上,该成本处理产品组件的平均成本–视通货膨胀,汇率和商品价格等许多因素而有很大差异。在专家估计单个项目的不确定性信息之后,决策者需要快速确定整个产品的分层数据结构中的成本不确定性。

我们建议只有包含数据和不确定性的整体可视化可以提供这种快速概览。这种可视化必须能够可视化与值属性关联的树数据结构。在与产品成本专家进行访谈之后,我们重点研究了满足此基本要求的流程图。但是,它们需要扩展以便直接合并不确定性信息。

我们研究了适用于流程图功能区以传达不确定性信息的三种可视化技术:颜色代码,渐变和边距。此外,我们设计了五种视觉方法来显示我们与可视化专家一起评估的流程图节点的不确定性。这些方法将不同的几何形状添加到节点,例如三角形,块或叉。节点的首选解决方案是添加叉子或填充块。关于功能区,我们提供了一项用户研究,其中涉及使用三种不同的可视化解决方案来解决不同的产品成本核算任务。尽管认为Gradient是显示不确定性的直观选择,但它产生的错误率最高。相反,颜色代码和边距取决于所执行的任务。

更新日期:2019-04-23
down
wechat
bug