Engineering insightful visualizations
Graphical abstract
Introduction
Many works have suggested that insight is the goal of visualization. As early as 1999, Card et al. [6] indicated insight as the purpose of visualization in the context of information visualization. Whereas, the panel discussion at IEEE VIS 2005 focused on “… the process of transforming data into insight.” in the context of a broader visualization scope inclusive of data and information visualization [12]. And, more recently, this goal also applies to visual analytics [31], [17], [7]. However, other visualization goals have been suggested: for example, understanding [17], knowledge [37], and decision making [38].
Visualization is broad in scope and has been used throughout history [35], [19]. It is often described as encompassing the three main subfields of Data (Scientific) Visualization, Information Visualization and more recently, Visual Analytics [27]. However, the distinction between these subfields has been a subject of debate [28], [33], [32], [36]. Although work is reported on modeling visualizations, for example, van Wijk [37], Keim [17] for visual analytics, and Chen et al. [8] who describe an information-theoretic framework; nevertheless, it has been observed that information visualization lacks a solid theoretical foundation [25], [39]. In general, many design guidelines have been considered in the literature, for example, Tufte [35], [34] and Ware [38], many of which in the latter are based on perception-related factors.
In this backdrop of visualization broadness, there is ample opportunity, interest and benefit in further considering the roles of understanding, knowledge, insight, decision-making, etc. The proposed work is also motivated by an engineering perspective influenced by how to quickly design and generate ‘good’ and ‘effective’ visualizations. These motivations are captured in the naming of this model, Engineering Insightful Visualizations (EIV) model, (and which is also the title of this paper); although it may be more apropos to consider it more as towards an engineerable visualization future.
The EIV model is a theoretical visualization model that is suitable for a guideline based engineering approach as well as generically and widely applicable to visualization and its subfields. It is based on investigating question–answer pairs and emphasizes understanding and knowledge acquisition achieved via insight and learning but which is impeded by confusion brought on by in-appropriateness, incoherence, anacolutha and non-sequiturs. These terms are technically defined within the model and are used in establishing analyses and guidelines by which to construct ‘better’ visualizations; that is, visualizations which are designed to facilitate increased insight leading to faster satisfactory understanding and knowledge acquisition. A visualization metric is developed that relates insight, learning and confusion with characteristics of how much and how fast understanding and knowledge are acquired. The model entails two connected processes. A visualization process generates a visualization as a media consisting of component parts (for example, multiple visualization techniques combined with imagery and/or audio). This process abstracts well-known and well-established visualization generation models and processes (for example visualization pipelines [20]). A human process models the perception, interpretation, understanding and knowledge acquisition when viewing a visualization. Several case studies drawn from the various subfields of visualization show the potential of the proposed model.
The EIV model captures the essence of visualization within the context of both visualization's broad scope and the vast variety of potential human visualization engagements via theoretical notations. The notations are designed to express both the construction and generation of visualizations as well as evaluation of the visualizations for improvement in terms of understanding and knowledge acquisition via insight and learning. The EIV model is designed to be compatible or overlap with existing works and thought has been given to its consistency with other related but non-visualization theories including learning theory and neural-science theory. Applications of the EIV model, in particular, the EIV models's visualization metric, may be used to identify less effective visualizations and provide guidelines to augment the visualizations in particular circumstances.
Section snippets
Related works
Insight as defined by Britannica Academic Edition (online) [11] refers to “…immediate and clear learning or understanding….” Five aspects of insight for visualization are identified by North [22]: complex, deep, qualitative, unexpected and relevant. Alternatively, two aspects are discussed in [7], spontaneous insight as a form of problem solving and knowledge-building insight as a form of learning, along with the authors’ hypothesis of the relationship between knowledge and insight. Insight is
Engineering insightful visualization (EIV) model
The proposed EIV model is developed in two parts; each of which is captured by a respective process. The first, Section 3.1, models the generation of a visualization as a media output (usually a picture or movie), herein referred to as the V process: the data model is firstly described, followed by visualization components and lastly the visualization mapping. The second, Section 3.2, models the human involvement wherein insight, learning, understanding, and knowledge acquisition occur via
Applications
Several case studies drawn from scientific visualization, Infographics and information visualization illustrate the applicability of the proposed model.
The first case study, a scientific visualization (Section 4.1), concentrates on two things. First, details about the input data set, visualization components, visualization mappings and visualization construction exemplify the V process and illustrate the applicability of this part of the proposed model. Second, directed insight and learning are
Conclusions
The theoretical Engineering Insightful Visualizations (EIV) model emphasizes understanding and knowledge acquisition as visualization goals which are achieved via insight and learning. As such, the EIV model comprehensively captures the long-time yet recently emphasized notion of insight from visualizations. More than this, the EIV model places insight within a larger and more comprehensive context of insight together with learning, grounded in the sense making process, in the context of
Trademarks
AVS/Express is a trademark of Advanced Visual Systems Inc.
Acknowledgements
James Life is thanked for his many helpful comments especially concerning the insight and learning parts of this work. Drs. R. Ravelo and C. Carrasco at The University of Texas at El Paso, Texas, USA, are thanked for making the crystal atom data set available as well as their past help regarding appropriate visualizations of the data set.
References (39)
- et al.
Would you prefer pie or cupcakes? Preferences for data visualization designs of professionals and laypeople in graphic design
J. Vis. Lang. Comput.
(2014) - R. Amar, J. Eagan, J. Stasko, Low-level components of analytic activity in information visualization, in: Proceedings...
- AVS Advanced Visual Systems. (Online). Available:...
- et al.
What makes a visualization memorable?
IEEE Trans. Vis. Comput. Graph.
(2013) - K. Brodlie, N.M. Noor, Visualization notations, models and taxonomies, in: Proceedings of Theory and Practice of...
- K. Brodlie, L.A. Carpenter, R.A. Earnshaw, R. Gallop, R. Hubbolt, A.M. Mumford, C.D. Osland, P. Quarendon, Eds.,...
- et al.
Readings in Information Visualization, Using Vision to Think
(1999) - et al.
Defining insight for visual analytics
IEEE Comput. Graph. Appl.
(2009) - et al.
An information-theoretic framework for visualization
IEEE Trans. Vis. Comput. Graph.
(2010) - et al.
Manyinsights: a visual analytics approach to supporting effective insight management
Tsinghua Sci. Technol.
(2013)
A systematic review on the practice of evaluating visualization
IEEE Trans. Vis. Comput. Graph.
Information visualization and visual data mining
IEEE Trans. Vis. Comput. Graph.
An algebraic process for visualization design
IEEE Trans. Vis. Comput. Graph.
Cited by (3)
‘Maddie is Online’: A Creative Learning Path to Digital Literacy for Young People
2023, Computers in the SchoolsOpen our visualization eyes, individualization: On Albrecht Dürer’s 1515 wood cut celestial charts
2020, Information VisualizationAR-based value-added visualization of infographic for enhancing learning performance
2017, Computer Applications in Engineering Education