A Description Framework for Data Visualizations in Enterprise Information Systems

Abstract

Data visualizations play a prominent role in enterprise information systems in various flavors. Traditional bar, line, or pie charts, or timelines, heat maps, geographical maps, dashboard gauges, and complex relationship mappings are examples of visualizations that are frequently used in business application scenarios. Despite their extensive use, however, there is only few theoretic reflection on how characteristics of data visualizations can be described on an abstract level independent from concrete graphical rendering. This is an obstacle when it comes to consciously reflecting about the use of visualizations for communicating information, because in order to gain a justified understanding of what "good" and "appropriate" visualizations for specific use cases are, at least a common terminology for characteristics of different visualization types is required. This paper introduces a description mechanism for conceptual and perceptual characteristics of data visualizations, which abstracts from concrete visual characteristics, but incorporates a joint notion of the underlying conceptual information displayed by visualizations, together with perceptual qualities of the way visualizations are cognitively processed by the human mind. The suggested solution describes each visualization type as a multidimensional abstract space, with specific scale characteristics attached to each of its axes.