Information Visualization Design Research Articles

Taking Mendeleyev’s correspondence course

Through an examination of the organizational principles behind a highly successful visual information design employing rich prospect – Mendeleyev’s periodic table of the elements – this paper provides a set of principles for designers interested in creating tools for use in association with rich-prospect browsing interfaces. These principles emphasise the value of simultaneous structures visualized across the full range of available information, and include: the establishment of multiple matrices for pattern-finding, compression and inclusion, provision of a legend or key to the system, and the creation of potentially productive negative spaces. Interface strategies based on Mendeleyevian principles should be particularly useful for accessing large bodies of digital documents such as those often found in the humanities, maximizing the preretrieval research potential of a collection.

Network Visualization by Semantic Substrates

Networks have remained a challenge for information visualization designers because of the complex issues of node and link layout coupled with the rich set of tasks that users present. This paper offers a strategy based on two principles: (1) layouts are based on user-defined semantic substrates, which are non-overlapping regions in which node placement is based on node attributes, (2) users interactively adjust sliders to control link visibility to limit clutter and thus ensure comprehensibility of source and destination. Scalability is further facilitated by user control of which nodes are visible. We illustrate our semantic substrates approach as implemented in NVSS 1.0 with legal precedent data for up to 1122 court cases in three regions with 7645 legal citations.

Statistical representations from popular texts for the ordinary citizen, 1889–1914

Statistical representations in the popular almanacs published at the end of the 19th century in the Habsburg Empire are an early prototype of visualizing statistical data for popular consumption and informing the public of an ethnically and linguistically differentiated society. The purpose of this article is to analyze these statistical representations using a combination of humanistic and visual analysis methods. The article focuses on the persuasive aspects (rhetoric) of diagrams, how they visualize knowledge and information (pragmatics), and the nature and structure of visual sense making (semantics). The form and content are related to the social context and materiality of images. These early popular diagrams predate functional design defined by positivism and science. They are interpreted as a historical practice of knowledge visualization in the context of a post-Enlightenment information regime (epitomizing rational government and the privileged position of statistics as instrument of increasing the efficiency and social control of the state). The modernist approach shapes the same information regime in information science discourse. The analysis—based on principles of information visualization and functional design requirements—shows that these examples possess a sound visual structure based on functional design criteria, and that they integrate cultural context without distorting the data. The advantages of naturalistic and culturally rich visualizations that enable ordinary citizens to acquire knowledge through leveraging simple visual reasoning skills, reliance on mental models, and narrative conventions are identified.

Knowledge Precepts for Design and Evaluation of Information Visualizations

The design and evaluation of most current information visualization systems descend from an emphasis on a user's ability to "unpack" the representations of data of interest and operate on them independently. Too often, successful decision-making and analysis are more a matter of serendipity and user experience than of intentional design and specific support for such tasks; although humans have considerable abilities in analyzing relationships from data, the utility of visualizations remains relatively variable across users, data sets, and domains. In this paper, we discuss the notion of analytic gaps, which represent obstacles faced by visualizations in facilitating higher-level analytic tasks, such as decision-making and learning. We discuss support for bridging these gaps, propose a framework for the design and evaluation of information visualization systems, and demonstrate its use.

Guest Editors’ Introduction: Special Section on InfoVis

HREE papers in this issue of TVCG are expanded versions of ones presented at InfoVis 2004. These examples of the cutting edge of information visualization research show the maturing of the field, using novel methods for designing, developing, and evaluating tools that help users solve real problems on large and complex data sets, rather than just creating interesting displays. “Knowledge Precepts for Design and Evaluation of Information Visualizations” by Robert A. Amar and John T. Stasko won the InfoVis 2004 Best Paper Award. In this paper, the authors argue that efforts to date to design effective visualization tools often fail because they concentrate mostly on presenting data, as opposed to supporting analysis. Limitations of existing systems are grouped into two gaps in the analysis process; the Worldview Gap consists of the difference between what is shown and what is needed to make decisions, while the Rationale Gap consists of the difference between relationships viewed and the certainty and utility of the relationship. Analysis of existing tools using this framework reveals many ways in which techniques can be improved to better support high-level analysis. “An Insight-Based Methodology for Evaluating Bioinformatics Visualizations” by Purvi Saraiya, Chris North, and Karen Duca presents a novel method for assessing and comparing the effectiveness of visualization tools, namely, the number and type of insights discovered. The authors define an insight as a unit of discovery and identify several attributes and categories of insights. They then report the results from controlled experiments used to evaluate five software tools for visual exploration of microarray data. The results shed many insights into what makes a visualization tool effective.

Aesthetics Versus Functionality

Information Visualization (2002) 1, 93 – 94. doi:10.1057/palgrave.ivs.9500018 This is the second issue of our young journal. This issue continues what we started in the first issue – to provide an interdisciplinary forum for the field of information visualization. To shape the interdisciplinary forum, we continue to seek original and stimulating articles on a variety of topics that in one way or another contribute to the advancement of information visualization and have the potential to make a significant impact on our research agenda in the field. Contributions to this issue include one commentary and five original research articles towards the fulfillment of this goal. Aaron Marcus in his commentary article addresses a number of challenges which stemmed from the design of advanced vehicle information displays. He outlines these challenges in the broader context of information visualization and emphasizes their implications for information visualization beyond vehicle information displays. His article is entitled ‘Information Visualization for Advanced Vehicle Displays.’ The first research article in this issue is authored by Colin Ware, Helen Purchase, Linda Colpoys and Matthew McGill on cognitive measurements of graph aesthetics. Graph drawing is concerned with ways to depict a network algorithmically. Graph drawing algorithms typically rely on a set of aesthetic heuristics – rules that algorithms aim to optimize quantitatively. However, there have been few empirical studies of these aesthetic heuristics in the relevant literature. The authors of this article introduce a methodology for evaluating the cognitive cost of graph aesthetics and apply it to the task of finding the shortest paths in spring layout graphs, one of the widely known graph drawing algorithms in information visualization. They provide cognitive cost estimates for three important parameters: the length of the path, continuity, and edge crossings. The number of branches emanating from nodes on the path is also found to be important. The second research article is by Archana Sangole and George K. Knopf on representing high-dimensional data in a spherical self-organizing feature map (SOM). SOMs have a unique position in the history of information visualization. Traditionally, SOMs are represented in the form of a flat two-dimensional map. Sangole and Knopf present a three-dimensional color-coded surface model. They illustrate the use of their model with a small synthetic test data set and a large environmental database. Robert Spence addresses a wide-ranging question in the third research article: How should we design information visualization interfaces that would enable users to navigate effectively in information space? He introduces the concept of sensitivity encoding as part of his design guideline for navigation in information spaces. Sensitivity encoding is a feature of a data display. The goal of sensitivity encoding is to provide ‘road signs’ to guide users through information space. The fourth article is strongly application-oriented. Mohamed Sammouda, Rachid Sammouda, Noboru Niki and Kiyoshi Mukai describe a liver cancer detection system. The system can be used to analyze digitized color images of tissue samples obtained using needle biopsy. The key to the system is a combination of an unsupervised segmentation algorithm and an analysis algorithm based on image quantization. The segmentation Information Visualization (2002) 1, 93 – 94 a 2002 Palgrave Macmillan Ltd. All rights reserved 1473 – 8716 $15.00

A graphical notation for the design of information visualizations

Visualizations are coherent, graphical expressions of complex information that enhance people's ability to communicate and reason about that information. Yet despite the importance of visualizations in helping people to understand and solve a wide variety of problems, there is a dearth of formal tools and methods for discussing, describing and designing them. This thesis develops and evaluates the Relational Visualization Notation, a graphical notation based on relational algebra for the specification of information visualizations.

The Application of Legibility Techniques to Enhance Information Visualizations

This paper shows how previous research into navigation through urban environments, which has emerged from the discipline of urban planning, can be adapted to enhance the design of information visualizations. The paper draws on Kevin Lynch's seminal work on the legibility of urban landscapes in order to propose a set of general techniques which can be applied to the task of information visualization. It describes a specific instantiation of these techniques called LEADS, a legibility system which post-processes the output of a range of existing visualization systems in order to enhance their legibility. The paper provides four examples of the application of LEADS to different information visualizations. Following this, it discusses experimental work, the conclusions of which provide some tentative support for the likely success of this approach. The outcomes of this work are both a recognition of the important relationship between the disciplines of urban planning and the design of information visualizations as well as more concrete algorithms to be used by the designers of such visualizations.