Reality-based Idioms : Designing Interfaces for Visual Data Analysis that Provide the Means for Familiar Interaction

Abstract

Due to the increasing amount of data, visual data analysis has become more and more important. Visual data analysis makes it possible to use humans’ mental capabilities to quickly recognize visual patterns and find relevant information and make sense out of them. However, humans’ mental resources are not only required for the actual knowledge generation process but also for the operation of the analysis tools. The interaction with visualizations as data analysis tools is the focus of this dissertation. The fewer resources the user has to spend for operating the tool the more resources are left to interpret the visual representation of the data and make sense out of it. The research approach underlying this dissertation is based on the theory of Reality-based Interaction. Our physical and social environment is full of affordances, rules, and constraints that we are aware of and that frame our thinking and interaction with people and objects. A reality-based interaction technique makes use of this by aligning the operation of a digital system with our physical and social experiences from the real world. The goal is to create user interfaces that seem familiar to the user. In this dissertation user interfaces for the visual analysis of data based on reality-based visualization and interaction techniques are referred to as Reality-based Idioms. Reality-based Idioms often make use of novel input and output technologies, which, compared to traditional desktop systems, leverage the users’ preexisting and entrenched knowledge about the physical and social world to a much greater extent. The technologies examined in this dissertation include multi-touch displays, deformable displays, tangible user interfaces, and head-mounted augmented reality displays. The applied approach reflects the duality of Human-computer Interaction as a research and design discipline. In the sense of a research discipline, the thesis presents the theoretical foundations that explain when a user interface is perceived as reality-based and thus as familiar and summarizes them in the Model for Reality-based Idiom Design. In the sense of a design discipline, the practical applicability of the model is illustrated by three domain situations. These focus on navigating visual information spaces, filtering large amounts of data and, analyzing multi-dimensional data. In order to address these domain situations, the dissertation presents five Reality-based Idioms. The evaluation of these design artifacts helps to clarify the benefits of reality-based interaction for visual data analysis. With this approach, the work makes an important contribution to closing the gap between visualization and interaction research and to pointing out new ways to facilitate the operation of tools for visual data analysis.