A Meta-modeling Approach to Take into Account Data Domain Characteristics and Relationships in Information Visualizations

Abstract

Visual explanations are powerful means to convey information to large audiences. However, the design of information visualizations is a complex task, because a lot of factors are involved (the audience profile, the data domain, etc.). The complexity of this task can lead to poor designs that could make users reach wrong conclusions from the visualized data. This work illustrates the process of identifying features that could make an information visualization confusing or even misleading with the goal of arranging them into a meta-model. The meta-model provides a powerful resource to automatically generate information visu-alizations and dashboards that take into account not only the input data, but also the audience’s characteristics, the available data domain knowledge and even the data context.