Visualization – Culture and Context
[1]

Anna Ursyn (bio)
University of Northern Colorado

Ebad Banissi
London South Bank University

Abstract

A framework for Visual Analytics, Information Visualization, and Knowledge Visualization domains emerges from looking at information-rich disciplines such as humanities, psychology, sociology, and business, not just science-rich disciplines. No studies encompass all aspects of visualization because every year brings about new application domains and evolving scientific disciplines. The theme chosen for the iV 2007 Conference is “Shifting focus to wider understanding and application.” In general terms, the drive can be seen as shifting from visual and presentational exploration to visual analytics. Selection of topics for this review has been aimed at bringing general information to attention of readers without delving into theoretical and technical solutions.

 

Fig.1. Joohyn Pyune, Seven Sorrows

1. Introduction

The first part of this review concerns the studies on visual aesthetics and criticism that combine information visualization techniques with the principles of creative design, support research on visual cognition, and broaden our awareness of knowledge. Further overviews refer to papers on such visualization domains as visual analytics (that applies analytical reasoning facilitated by interactive visual interfaces – a combination of computational and visual methods in exploration process, Schulz and Schumann, 2006), knowledge visualization (that makes easier the creation and transfer of knowledge between people with the use of visual representations, Burkhard, 2005), information visualization (that uses computer supported, interactive visual representations of abstract data structures to amplify cognition, Shneiderman, 1996), and web mining (that enables the users to see patterns in the web data, http://en.wikipedia.org/wiki/Web_Mining). The overview includes also educational applications of visualization concepts and technologies. This review is illustrated with the selected examples of artwork presented at the Symposium and Gallery of Digital Art (D-Art).

 

Fig. 2. LIQuin Tan, Lava Body III-1

 

2. Aesthetic issues in information visualization

This part of the review presents explorations about visualization aesthetics and criticism: how information aesthetics influence technical implementations, thus being useful in choosing best techniques for a particular visualization; how identifying dimensions of visual aesthetics may allow the interface designer to evaluate and adapt the tools for the needs of the user; how artistic components fulfill both the aesthetic and technical requirements in effective visualization techniques; why visualization criticism can bridge the gap between design, art, and technical/pragmatic visualization; and what is the positive role and purpose of aesthetics in the design of data visualization techniques.

A model of information aesthetics that combines information visualization techniques with the principles of creative design has been proposed by Andrea Lau and Andrew Vande Moere (2007, pp. 87-92). It considers the context in which the data should be interpreted, rather than subjective judgment. The authors refer the concept of aesthetics to the degree of artistic influence on the visualization technique and the amount of interpretative engagement they facilitate. A model proposed by the authors is focused on aesthetics as the artistic influence on technical implementations and is seen as useful for designers in choosing best techniques for a particular visualization.

Effective visualizations are supported by insights from visual cognition and perception research, as well as taxonomies that match data types to the most effective mapping technique. The relationship of information aesthetics with different visualization techniques can be mapped according to three factors: data, aesthetics, and interaction. Lau and Vande Moere (2007) designed a model for information aesthetics where several existing visual representations of abstract data have been aligned against two axes, one focusing on the data (that span from extrinsic to intrinsic) and another one representing the mapping focus (that extend from the direct to interpretative applications). Thus, the extremes in this model include information visualization technique at the direct – intrinsic corner, while visualization art has been placed at the interpretive – extrinsic one. “Visualization techniques with intrinsic data focus aim to facilitate insight into data by employing cognitively effective visual mapping,” while “those with the extrinsic data focus facilitate the communication of meaning that is related to or underlies the data set” (Lau & Vande Moere, 2007, p. 90).

Dimensions of visual aesthetics (namely color, form, spatial organization, motion, depth, and the human body) and their specific characteristics called esthetic primitives are capable of evoking an aesthetic experience (Peters, 2007, pp. 316-325). According to Peters, the role of images (pictures, photographs, or drawings) is underrepresented in the literature. Identifying aesthetic dimensions may allow the interface designer to evaluate the aesthetic qualities of an image and adapt the tools for the needs of the user. Both cognitive neuroscience and artistic principles give us clues to define dimensions of visual aesthetics. Research results suggest that the foundation of aesthetics is in cognitive neuroscience. Human visual system is organized in modular system with attributes, such as color or motion, in different, specialized subdivisions of the brain. Some hold that aesthetic preferences correlate with specific brain structures and processing systems, so there is a functional specialization in aesthetics and the basic dimensions of visual aesthetics are equal to the different modularities of the human visual system.

Six dimensions of visual aesthetics that were derived by the author (Peters, 2007) are: color, form, spatial organization, motion, depth, and the human body. In color modularity, the selected properties that appeal to the user is: the choice of only a few strong colors, the use of complementary contrast, and exploitation of dynamic range. Clarity of form and the use of silhouettes have been selected as components of aesthetic images, along with clarity of spatial organization, application of the golden mean, the use of textures, patterns, rhythm, repetition, and variation. Motion, discussed as expressive motion symbols such as blur or depiction of distinct motion phases, and depth, defined by linear perspective, the contrast between sharpness and unsharpness, and distribution of line or shadow, are other essential aesthetic cues. In human body, the concept of principal axes is considered important.

Artistic component is required in effective visualization techniques, to fulfill both the aesthetic and technical requirements (Erbacher, 2007, pp. 623-630). However, computer scientists may not receive the level of training in design that artists and architects do. Therefore, data visualization research requires the integration of many disparate domains, including computer science, mathematics, statistics, art, architecture, cognitive psychology, and cooperation with the domain experts as the users for whom the visualizations are being designed. The success of visualization derives from its reliance on human perception. Visualization maintains the user in the loop, allowing for intuition and expertise to take part in the analysis process. Thus, visualization represents both data in its raw form as well as the computed values, often as statistical or mathematical results. Data sets can be huge, with hundreds or thousands of parameters. Sometimes it is not known what the analyst should be looking for such a data sets. Apart from the automated techniques, the human analyst must direct the exploration and analysis processes, interpret and correlate the huge amounts of raw data with large amounts of computed data (Erbacher, 2007).

Nick Cawthon and Andrew Wande Moere (2007, pp. 637-643) indicated in their research that aesthetics plays a positive role and purpose in the design of data visualization techniques. The authors examine the effect of the data visualization aesthetics on specific measures of usability (that were originally defined for website evaluation) namely, the correlation between perceived aesthetics, task abandonment (a measure of usability primarily referenced within the field of web analytics, a study of user behavior on the Internet), and erroneous response times (length of time taken by a participant who generates an incorrect answer). In an online survey, participants performed aesthetic ranking of 11 different visualization techniques and provided answers telling about task performance. The research results indicate that aesthetics has an effect on extending the latency of task abandonment and duration of erroneous response time: the most aesthetic data visualization technique, the Sunburst type, performs relatively high in metrics of effectiveness, rate of task abandonment, and latency of erroneous response. These factors correlate with user’s patience, the duration in which interaction occurs before either completion or abandonment.

According to Robert Kosara (2007, pp. 631-636), visualization criticism can be considered the missing link between information visualization and art. Information visualization requires interdisciplinary, integrated approach and a strong interchange of ideas. The author sees two cultures in visualization: very technical, analysis-oriented work, and artistic pieces. He holds that people in computer science with no background in art or design practice mostly pragmatic visualization, while artists and designers often work on visualization without much knowledge in computer science. In pragmatic visualization, visual efficiency is a key criterion for work in visualization, with images that convey the data quickly and effortlessly. User studies are conducted to measure the speed and accuracy of users and to compare different methods. The goal of artistic visualization is usually to communicate a concern, rather than to show data that is used as the raw material.

Fig.3. Helen Golden, Idea of Reflected Surfaces

Visualization criticism, which can be applied to both artistic and pragmatic visualization, can be seen as bridging the gap between design, art, and technical/pragmatic visualization (Kosara, 2007). A foundation theory of visualization is still missing; there is very little discussion of approaches, with many techniques developed ad hoc or as incremental improvements. The use of visualization criticism is a possible path towards developing a theory and a language that are largely missing in visualization. Many visualization researchers use criticism in their visualization classes.

Kosara considers several criteria as a minimal set of requirements for any visualization: it is based on (non-visual) data, it produces an image, and the result is readable and recognizable. Transformation of data into a visual shape does not imply readability. Visual mapping from data may not provide information about it, like in some music visualizations, informative art, or ambient visualizations.

Critical thinking may connect the technical approaches to pragmatic visualization with philosophy and artistic visualization. Criticism is an important part of this process. One aesthetic criterion is the sublime that can be understood as that which inspires awe, grandeur, and evokes a deep emotional and/or intellectual response. The user friendliness is considered the opposite, and is a central concept in computer science, where visualization techniques need to be evaluated in user studies, and by removing any sublimity they are designed to foster immediate understanding. The author proposes several rules to guide the process of visualization criticism: the neutral voice aimed to discuss the work, not the researcher; statements based on facts that can be independently checked; no self-promotion; and a clear goal in stating alternate solutions, not just criticizing a work for its shortcomings. The ultimate goal of visualization criticism is to provide building blocks for a theory of visualization, when theory and practice are used to develop, evaluate, and validate each other. According to Kosara (2007), ideas common in the arts can be appropriated and modeled into a modus operandi acceptable for a scientific discipline.

3. Visual Analytics

An emerging framework for information visualization involves analytic inquiries facilitated by interactive visual interfaces. Selected summaries include: a note about Coordinated and Multiple Views (CMV) in Exploratory Visualization; a multi-dimensional analysis of user-based managerial visualization methods; information about the geoanalytics visualization toolkit that provides interaction code for a mixture of technologies from the three visualization fields: information visualization, geovisualization, and scientific visualization; and description of a concept-mapping tool is a search interface that creates interactive concept maps for user’s queries.

The area of Coordinated and Multiple Views (CMV) in Exploratory Visualization (EV) is developing over the past fifteen years as a tool for Visual Analytics (Roberts, 2007, pp. 61-71). The user interacts with the data to both formulate the problem and concurrently solve it. CMV is a specific exploratory visualization technique that enables users to explore and understand their data better if they interact with information, view it through different representations, “compare visualizations generated from multiple different datasets, aggregate and mine the data, perhaps fuse data from multiple different datasets to generate new information, and be able to easily roll back to a previous incarnation” (Roberts, 2007, p. 61). According to Jonathan Roberts (2007), seven fundamental areas of CMV are: data processing and preparation; exploration techniques; coordination and control; tools and infrastructure; human interface; and usability and perception.

 

A study on visual methods for management provides taxonomy of user-based managerial visualization methods. In their research study on empirical classification of visual methods for management, Martin J. Eppler and Ken Platts (2007, pp. 335-341) conducted picture-sorting experiments with managers and students. Participants grouped images of visualization methods by similarity. Usually, managers do not know visualization methods and use pie-, bar- and line-charts only. Students (engineering and manufacturing management) and junior and senior managers sorted 30 picture cards showing visualization-based methods. Average cluster analysis was then performed and presented as a dendrogram of 30 visual methods, and then the multi-dimensional (MDS) scaling analysis was performed, in 2 and 3 dimensions based on similarity ratings. There were significant differences in the way that men and women group the visual methods. Discussion of the results can be useful in future classification systems in the area of managerial visualization methods.

The Geoanalytics visualization toolkit GAV presented by Mikael Jern and Johan Franzen (2007, pp. 511-518) allows for dynamically exploring time-varying, geographically referenced and multivariate attribute data. GAV includes components that support a mixture of technologies from the three visualization fields: information visualization, geovisualization, and scientific visualization. GAV provides interaction code for event handling, brushing, zooming, and drag-and-drop, without a necessity to write a code.

A concept mapping tool, Visual Concept Explorer (VCE) developed by Xia Lin, Yen Bui & Dongming Zhang (2007, pp. 476-481) allows for visualization of knowledge structures. VCE can be used as a search interface in a practical environment, as it creates dynamic concept maps in real time for user’s queries, which interact with concepts and documents and make available explicit and implicit knowledge structures. It combines knowledge structures automatically extracted from texts with knowledge structures created by human experts. VCE demonstrates the promise of synergy between information visualization and knowledge visualization.

 

Fig. 4. Hans Dehlinger, WUELIZ_02

 

4. Visual Analytics  – Web Visualization

Environments created for visualization and analytics of the web search results include interactive maps created with the visual mashup approach and semantic models.

With the use of interactive tag maps, tag clouds, and a mashup approach, Aidan Slingsby, Jason Dykes, Jo Wood, Keith Clarke (2007, pp. 497-504) can prototype a set of techniques to test different solutions and identify issues for further research on of large spatio-temporal datasets. Tag clouds and tag maps represent geographically referenced text. Tags are free form text labels that are independent of controlled vocabulary. They are widely employed for labeling digital content, such as photographs (Flickr, www.flickr. com), videoclips (YouTube, www.youtube.com), and WWW bookmarks (www.del.icio.us). Tag clouds are a visualization technique that summarizes collections of words other than tags (www.tagcrowd.com). Tag maps are tag clouds grounded in real geographical space. Mashups often use mapping or graphical technologies as the basis for integration. A mashup approach uses combining a set of freely available network-friendly technologies that use de facto data standards with published APIs, for rapidly prototyping the techniques.

A mashup approach has enabled the authors (Slingsby et al., 2007) to rapidly prototype a set of techniques, to test different solutions and identify issues for further research. The interactive tag map and tag cloud techniques and the rapid prototyping method provide spatial and aspatial views for exploring large structured spatio-temporal data sets by providing overviews and filtering by text and geography.

Creating a combination of abstract data types that do not have an obvious spatial mapping is considered a key challenge in the field of Information Visualization (Spoerri, 2007a, pp. 216-221). With a visual mashup of text and media search results, users view diverse sources of data in an integrated manner, when geographical meta-data is not available or advisable to use to combine the data sources. Anselm Spoerri (2007a,b, 2007) developed Crystal tools that can visualize the overlap between any fuzzy sets: the searchCrystal toolset is used to visualize web, image, video, news, blog, and tagging search results in a single integrated display; MetaCrystal to visualize the overlap between multiple engines searching for the same data type; and InfoCrystal makes it possible to formulate and visualize Boolean as well as vector-based queries in the same visualization. searchCrystal enables users to create a ‘multi-media’ snapshot of a person, company, or topic of interest.

Imago, an integrated prototyping, evaluation and transitioning environment for information visualization, presented by Rudi Vernik, G. Stewart Von Itzstein, and Alain Bouchard (2007, pp. 17-22) provides a distributed environment that supports the prototyping, evaluation, and transitioning of information visualization solutions to users. The semantic model of contextual and visualization knowledge can store various semantic relationships between the reference model for visualization concepts, evaluation results, and instrumentation data.

According to Edward Suvanaphen and Jonathan C. Roberts  (2007, pp. 238-244), about one-third of searches that are performed on the web require the user to initiate subsequent searches to acquire new information that in turn leads to new ideas and directions. This process changes the query terms and also the nature of information retrieval task itself. Visualizing evolving searches with EvoBerry environment enables the users to find, view, and manage data produced from their searches, along with methods to visualize additional search result information (such as length of page or file type), manage the user’s session and browsing history, compare result sets, and store and bookmark items for future reference.

A self-organizing meta-search engine makes possible yet another web search result visualization based on a semantic map (Hamdi, 2007, pp. 222-227). Information retrieval mechanisms provided by Internet Web software are usually based on either keyword search (e.g., Google and Yahoo) or hypertext browsing (e.g., Internet Explorer and Netscape). The self-organizing meta-search engine SOMSE replaces long lists of ranked documents, as it queries popular search engines (Google, Yahoo, and Msn) as information filters, and then returns a single list of up to 60 documents as a two dimensional self-organizing semantic map with the visible underlying structures of the document space. In few seconds, SOMSE generates correct clusters with meaningful short names that could improve users’ browsing efficiency through search results. It is an information customization system that combines meta-search and unsupervised learning. The Kohonen Feature Map is used to construct a self-organizing semantic map as a browsing aid (Hamdi, 2007).

The Hierarchical Visualization System (HVS), a general framework, implemented in Java, that provides a synchronized, multiple view environment for visualizing, exploring, and managing large hierarchies has been developed by Keith Andrews, Werner Putz, and Alexander Nussbaumer (2007, pp. 257-262). HVS has been also designed to provide a platform for the empirical evaluation and comparison of different hierarchy browsers. Eleven hierarchy browsers have so far been implemented within HVS, including: traditional tree views, the classic Walker tree layout, information pyramids, treemaps, a hyperbolic browser, sunburst, and cone trees. Information slices, a radial node-link layout, a botanic visualization, and a recursive voronoi browser will be implemented soon.

Fig. 5. Bogdan Soban, Reintegration 1

 

5. Knowledge Visualization: Theory and Investigation

Visualization of knowledge domains and domain mapping attracts much attention of the researchers who approach both the theoretical and technical problems by exploiting the experience gained in other domains such as semiotics, advertising, architectural design, and movies. Much investigation has been launched into the use of color in visualization. A thorough analysis of techniques in information visualization often involves application of graph theory; inquiry has been made about comprehension of diagrams and graph drawing aesthetics.

Stefan Bertschi (2007, pp. 342-347) proposes a deconstructivist approach to metaphor, meaning and perception. The first symposium on knowledge visualization took place at the 9th International Conference on Information Visualization (iV05, London) and the D.Sc. thesis by Remo Burkhard (2005) was the first major scholarly work on knowledge visualization. Burkhard stated there that knowledge visualization aims to improve the transfer and the creation of knowledge by giving people richer means to express what they know. Bertschi and Bubenhofer (2005) examined the metaphorical face of knowledge visualization and applied criticism on a missing theoretical foundation that needs to be built before establishing visualization science as a scientific discipline. According to Bertschi, “a deconstructivist approach may facilitate the design as well as the evaluation of knowledge visualization tasks” (2007, p. 346). With pictorial and linguistic visualization being complementary parts of communication, knowledge has to be re-constructed by each individual through communication and interaction with explicit verbal or visual information with the use of existing structures and patterns.

In knowledge visualization, visual metaphors are used to carry complex concepts and visual storytelling to disseminate knowledge. Because knowledge visualization is a process of construction and reconstruction, the only reasonable approach to evaluate its mechanisms and effectiveness has to be deconstructivist. As Bertschi put it, “every deconstructivist approach can be seen as some kind of depiction or visualization, and every visualization should be seen from such a perspective” (2007, p. 343). The task to be done is to combine investigating principles of knowledge visualization and the use of metaphors by exploring their semiotic and symbolic components. Thus Bertschi concludes that metaphor (a form of thought playing fundamental role in acquisition of knowledge), meaning (carried by language and derived from things by representation) and perception  (“a set of processes by which we recognize, organize, and make sense of stimuli in our environment” Burkhard, 2005, p. 40) are not sufficiently elaborated in knowledge visualization. However, according to Burkhard, visualizations need to be customized to transfer not only facts but also meanings and insights, so the recipients with different backgrounds can re-construct knowledge as intended by the sender.

Visualization of domain mapping serves for discovery, understanding, communication, and education. Peter A. Hook (2007, pp. 442-446) examines purposes, history, parallels with cartography, and applications of domain maps and knowledge domain visualization. Domain mapping is the graphic rendering of bibliometric data designed to provide a global view, structural details, and the salient characteristics of a particular domain. Visualization of domain mapping may facilitate creation of associations between concepts and spatial metaphors according to the neural theory of metaphor and knowledge. And also can be applied as front-ends to a body of literature, online learning environments, and digital libraries.

Fig. 6. Leslie Nobler Farber, Dorothy’s Wings

Ralph Lengler recommends learning from advertising research for information visualization (2007, pp. 382-392) because advertising is the oldest visualization field that addresses head and heart. Unlike visualization, advertising amplifies emotion, not cognition, as visualization expands processing capacity of the decision maker, and advertising induces ‘somatic’ decision markers. The measuring of the heart response has recently been based on data, derived from cognitive psychology and neuroscience, about how meaning is created and memories retained when looking at visuals and how interpretation of visuals affects persuasion over the emotional route.

According to Lengler (2007), research on advertising disclosed three recurrent themes that refer to the interplay of cognition and effect in the interpretation of visuals: ‘likeability’ that makes the best predictor of advertisement effectiveness; emotional advertising that frames perception and influences persuasion; and a ‘co-authoring’ effect that involves the beholder in meaning creation, thus effectively making him a co-author. Advertisers want to be convincing so they use a lecture approach (logic, reason, and rhetoric) to persuade the consumer, and also the story/drama approach to evoke feelings and create emotions. They want to be remembered at purchase time.

A longstanding debate refers to the question about what is more important – the message content (as advocates David Ogilvy) or its creative execution (as promoted by Leo Burnett), it means, whether what you say is more important than how you say it.

Neuroscientific inquiries led to conclusion, described by Damasio in 1994 and 1999, that while neocortex performs reasoning and analysis, it is still wired up through the old biological brain, so “emotions and feelings will always be formed precognitively and preattentively before any information processing takes place” (Lengler, 2007, p. 384). Thus, emotion leads to action while reason leads to conclusion. Emotional somatic markers endure when rational arguments fade, which explains the pervasive use of emotional advertising.

Implications for Visualization Science drawn from the advertisement study results include likeability, co-authorship, and the role of ‘love.’ In information visualization literature, we find similar characteristics constituting likeability of visual representations, namely visual perspective and information context. But the framing of the message – what people hear – may unconsciously moderate the brain how to process visual information. Recently, the balance has been shifted from cognitive textual to the affective visual. When compared with advertising, visual representations in information visualization domain correspond well in such dimensions as likeability – ease of information assessment and comparison, and visual perspective (namely, interactivity and depth of field). Information visualization should master also vividness resulting from “arranging a harmonious ensemble” (Lengler, p.391).

In the domain of architectural visualization, evaluation of the potential of the augmented reality in urban environment design and the computer-generated forms of architectural representations revealed controversial opinions. Experimental evaluation of an augmented reality technology (ART) against traditional wood block method (TWB) for urban design and planning, conducted by Xiangyu Wang, Rui Chen, Yan Gong, and Yi-Ting Hsien (2007, pp. 567-572) revealed some technical difficulties in the ART system that introduce negative performance factors. The advantages of TWB over ART in urban design are: the model control and visual perception. The users postulated a mixed usage of both tools. The ART system that uses virtual representations of objects could be further developed into distance collaboration tool by networking remote participants. The collaboration tool can be used along with video and audio conferencing software to enhance the effectiveness (Wang et al., 2007). 

Nada Bates-Brkljac (2007, pp. 348-353) investigates perception, understanding, assessment, and subjective judgments about architectural design ideas when experts and non-experts responded to traditional and computer-generated forms of architectural representations. Architects (as experts) and member of the public (as non-experts) apparently perceive and understand visual representations differently. However, the author finds strong resemblance between professionals and lay people in perceptual and cognitive responses to traditional and computer-generated forms of architectural representations. Architects’ responses varied depending on the age and length of work experience.

The public and members of planning committees receive information through visual representations. Perception, understanding, assessment, and subjective judgments depend on understanding and knowledge exchange through visual communication. The Bates-Brkljac study (2007, pp. 348-353) investigated people’s perceptual and cognitive responses to traditional and computer-generated forms of architectural representations. It examined an impact of the use of computer technologies on professional relationships across interest groups in architecture, especially the communicative role and effects of computer-generated representations. Concepts under discussion were accuracy, abstraction (level of detail), and realism. Participants provided evaluative responses to questionnaires (eighteen bipolar Likert-like questions on a seven point scale) about these criteria, along with their individual views. The author’s conclusions are that “there is no such thing as accurate and credible architectural representation; instead it is evident that representations enter people’s perceptions through a complex interaction of choice, constraint, and visual literacy” (Bates-Brkljac, 2007, p. 352); also, that a credible representation should bear some resemblance to the photographs, and also that it should be easy to understand and avoid architectural ‘visual jargon.’

Ismo Rakkolainen (2007, pp. 935-942) investigates how feasible are Star WarsTM Mid-air displays. Some existing in reality displays match with some mid-air displays seen in the Star Wars movies. However, no current technology is yet on a mature Star Wars level in all aspects. Ideas that were first presented in science fiction books, comics or movies (such as invisible clocks, walking through the wall, impression of a 3D image floating in free space, holographic and mid-air displays) are realizable, to some extent, with novel technologies. The closest to Star Wars mid-air displays is the immaterial FogScreen technology. The classical mid-air display, the ’hologram’ of Princess Leia is trivial to implement with side-projection. Bottom-projection (where images are seen above a platform and the projection comes from below the platform) would be problematic and large surrounding mid-air displays covering the whole room (like the stellar map shown by Obi-Wan Kenobi in “Episode II”) are not possible with it.

Rakkolainen holds that since the eye only captures 2D images on the retina, 3D perception comes from a variety of cues that imply depth in the scene. Images, objects, and user interfaces that seem to float in mid-air can be generated in a variety of ways by artificially recreating the effects of depth cues from natural viewing. Potential mid-air display technologies include stereoscopic and autostereoscopic displays (that provide slightly different images for the left and right eye), head-mounted displays (often used in virtual and augmented reality), volumetric displays (where a 3D image of an object can be seen within a volume from arbitrary viewpoints without any eyewear), holography, laser and other projections, floating images (with no visible screens but configurations of or half mirrors), transparent projection screens (where the viewer sees only the areas of the screen  where objects and light is projected), and immaterial displays (with some medium such as air, dust, gaseous or liquid particles, e.g., water, smoke, or the FogScreen technology).  Mid-air displays could hit the mainstream and revolutionize application areas such as CAD, data visualization, digital signage, tele-presence, simulation, and entertainment. However, the user’s attitude and not technical high fidelity might be the most important factor, because a good novel can be more ‘immersive’ than poor content with a high-end virtual reality system (Rakkolainen, 2007).

Much attention is paid to the use of color in visualization. Samuel Silva, Joaquim Madeira, and Beatriz Sousa Santos (2007, pp. 943-950) review the tools and methods used to manipulate the color scale in visualization. Data is often mapped onto a visual structure by using color mapping. However, some visualization practitioners use color without asking if the information they want to depict is still clearly understood. Guidelines on color use in visualization and the desired properties for color scales, developed by researchers, help the users along the process of color scale selection according, for example, to the type of data and task to be performed. The desired properties for color scales include the order the colors chosen to represent the values must be perceived as having the same order as the values, e.g., cold and warm colors for a temperature scale), uniformity and representative distance (colors representing values which equally differ should seem equally different, e.g., in flow information, complementary colors can be used to represent flow in opposite directions and almost similar colors can show flow in the same direction), boundaries (the color scale must be able to represent continuous scale without boundaries), rows and columns principle (rows and columns with constant value of one variable must have constant hue, saturation, or brightness).

As Silva and co-authors (2007) put it, univariate color scales may form continuous path (with adjacent colors similar to one another) or may contain discontinuities. Color model components include a gray scale (that maps the value of a scalar to brightness), and a spectrum scale (also known a rainbow scale, which holds the saturation and brightness constant while letting hue vary through its entire range). Redundant color scales may vary in both luminance and hue, to accurately represent both metric and surface properties. Double-ended color scales are created by joining two monotonically increasing scales at a common end point, e.g., from red to gray to blue, to visually represent high, low, and middle values clearly. Multivariate color scales map two or more data variables to a single color (for example, Landsat ‘false color’ images). In a color scale selection, one has to consider the content and type of data to be represented, so the most striking features of the image reflect the most important features of the data. By observing conventions typical of an audience (e.g., some place blue/violet color of a spectrum scale at the low end, in order of increasing wavelength, while others place it at the high end, in order of increasing frequency) and cultural connotations (e.g., high temperatures represented in red, low in blue color) it is possible to reduce cognitive load on the viewer. Visualization type is important, with viewers using shading cues to judge the shape in 3D visualizations. Research and user studies provide information on human color vision and evaluate tools and methods offered to interactively manipulate the color scale in exploration of data sets and selecting color scales (e.g., the PRAVDAColor tool or ColorBrewer) (Silva et al., 2007).

Wibke Weber (2007, pp. 354-359) visualizes texts by colors to explore what colors tell about a text. Text visualization include a color code that displayed word classes: noun, verb, adjective, determiner, particle, conjunction, and interjection. The colored words provided information about text genre (scientific texts or fictional narratives), sentence structure, and writing style. A sentence with the same message can have different text-images depending on its writing style. When color-coded, scientific and narrative texts show different color patterns, fictional texts having more verbs and thus brighter color pattern. Existing visualizations served for text mining by browsing textual data, illustrating patterns in data, localizing specific topics, or providing text summaries and formal analyses.

Techniques in information visualization often involve application of graph theory. Kamaran Fathulla and Andrew Basden (2007, pp. 951-956) review definitions of diagrams and propose a framework for a good understanding of a diagram. Writers who build definitions of a diagram encounter problems related to the variety of diagram types, their meaningful dynamics, and handling changes in diagrams, all of this in the context of semantically mixed diagrams. As for variety of diagram types, diagrams often contain a mix of different types or classes of diagrams, e.g., surface coverage and bar charts. Clay tablets from before 2,500 years expressed boundaries, groupings, and routes. Handling changes may present problems because one has to obey rules (e.g., in a contour map lines cannot overlap, in a line-and-box diagram lines cannot connect with a box), sometimes rules need to be relaxed (e.g., when drawing part of the object to represent a whole), rules are not arbitrary (e.g., when certain materials afford certain meanings), and diagrams present variations in different relationship concepts. According to Fathulla and Basden (2007), definitions developed by several authors emphasized various issues, e.g., Peirce (1931) put an emphasis on what we do with diagrams, Gombrich (1966) – on interpretation and new meaning, Bertin (1983) – structure of diagrams in graphical terminology, Ittelson (1996) – interpretation and communicative intent, Knoespel (2001) – what we do with diagrams, and Engelhardt (2002) – cognitive structure of diagrams. A proposed framework for a good understanding of a diagram based on Symbolic and Spatial Mapping advices separating Symbolic from Spatial but allowing for their Mapping. The term SySpM denotes a particular collection of Sy and a distinct collection of Sp and a distinct Mapping between the two Sy and Sp collections. The symbolic aspect Sy gives precision without ambiguity when similar spatial phenomena are in many types of diagrams. For example, in the London Underground Map circles and lines (Sp) express a range of Sy meanings diagram (Fathulla & Basden, 2007).

While evaluating the comprehension of Euler diagrams, Florence Benoy and Peter Rodgers (2007, pp. 771-778) investigated the importance of various Euler diagram aesthetic criteria. Euler diagrams are represented by interlinking sets, often with dots or graphs present in the diagram to indicate which set particular items belong to. Venn diagrams are a special case of Euler diagrams, where every possible zones (areas produced by the intersections of circular contours) are present. Euler diagrams are becoming a widely used technique in information visualization. Research in automatically laying out Euler diagrams is in early stage.

The research on layout criteria that can help with the comprehension of Euler diagrams has been done to support decisions concerning the metrics that mandate automated layout of Euler diagrams (Benoy & Rodgers, 2007, pp. 771-778). The choice of the criteria was based on the findings of research into graph drawing aesthetics. Three criteria have been chosen: Contour Jaggedness that relates to the continuousness of the contour lines, Zone Area Inequality that relates to the relative sizes of the zone areas, and Edge Closeness that relates to the closeness of lines from different contours. These criteria have been found important with regard to the diagram layout. Interactions between criteria become stronger in complex diagrams.

Fig. 7. John Labadie, Molecular Modeling Series

6. Information- and Knowledge Visualization – Applications

A few selected presentations provide only a partial view of the vast and growing resources developed every year and made available for users. Themes selected here include presentations about visual methods for management using the user-based managerial visualization methods; virtual and real-world environments to represent, navigate, and interact with complex data; application of virtual and simulation systems to garment production; and examination of rendering devices.

Information visualization approaches to retail space management (VisMT) let the users integrate interactive store plans with retail data with a visual user interface. A novel approach to retail space management analysis, proposed by Mikael Jern (2007, pp. 109-116), is a synergy of 3D layout store floor plans, common interactive information visualization methods, and multiple linked views. A visual user interface VisMT integrates familiar visualization representations and a 3D interactive layout of store floor plans with retail data sources. Retailers gain precise understanding of each store’s layout in relation to its capacity and performance.

Wolfgang Kienreich, Mario Zechner, and Vedran Sabol (2007, pp. 363-368) developed comprehensive astronomical visualization for a multimedia encyclopedia where users can interactively navigate in space due to integration of planetarium visualization and a virtual theater. Multimedia encyclopedias are capable of displaying complex, 3D visualization in real-time, enabling the integration of planetarium and a virtual theater. Planetarium visualization has been integrated into the “Brockhaus Multimedial” encyclopedia and the authors discuss design and implementation considerations, the astronomical entities, applications used, combinations of metaphors and real-world models implemented into this project. User interaction, based on a concept of a virtual observer located at an arbitrary position within a solar system, involves interaction modes: the ‘orbital mode’ (that places the observer in orbit around the entity like a planet or moon), the ‘look mode’ (the observer’s location is kept constant), and the ‘overview mode’ (with the sun placed at the center). By providing users with a virtual observer, the authors caused that a spatial dimension could be exploited for navigation and interaction.

A multi-dimensional visualization technique, based on a concept of Data Forest, has been designed by Ronald Jamieson and Vassil Alexandrov (2007, pp. 293-298) to be used with virtual reality (VR) as a presentation method. It uses data trees to represent the data. Users can walk and/or navigate through the forest of data trees rendered in virtual environment that represent the complex data in an affective manner. The Data Forest concept is based on the standard visualization approach of having a pipeline that consists of a data source at the start of a pipeline and rendering 3D objects to a display at the end. By applying a filter or algorithm to the data source, it is possible to create a series of data structures to store the data. Then using these data structure the authors generate a forest of data trees rendered in virtual environment (VE) that represent the complex data in an affective manner. Any one data tree can visualize eight different parameters: position of the tree within VE, height, width, color, and transparency of the trunk, size, color, and transparency of the crown. Using the immersive VR system, authors develop an interactive VE to represent different level/layers of abstraction of the data by creating different Data Forests for further data mining. Specific functionalities that can be added to the system include: changing simulation mode from static to a dynamic or simulation mode; changing the scale of the forest to improve overview, interaction, or exploration; exploring underlying data; predicting functions for further trees; and creating collaborative sessions by networking virtual environments together, representing remote users as avatars (to be mapped to the remote user’s position, orientation and gestures) and communicating with them via audio and video links.

The cost- and time-effective representations of lighting environment can be achieved by combining a sequence of differently exposed images into one image. Instant realistically rendered representations and extreme special effects are time consuming, difficult to use, complex, and expensive. The emerging technology, the High Dynamic Range Image (HDRI) developed by Ahmad Rafi, Musstanser Tinauli, and Mohd Izani (2007, pp. 877-882) is a combination of multiple images with different exposures, so it captures the dark and the bright regions that are present in a scene under observation. Applications of HDRI include cost and time effective solutions in the field of HDRI-based lighting and environment (for example, capturing different light intensities at a designated surrounding) and architectural visualization (for example, a 3D model of helicopter imposed on the existing real environment to compare and discuss certain design factors).

Funda Durupinar and Ugur Güdükbay (2007, pp. 862-867) present a system where, in a 3D graphic environment, the virtual garment construction, design, and simulation processes can be performed through automatic pattern generation, posterior correction, and seaming, and then fitting on virtual mannequins as if in a real life tailor’s workshop. They also present rendering alternatives for the visualization of knitted and woven fabric.

According to Lin Hsin Hsin (2007, pp. 845-849), a handheld device, a two-button mechanical mouse, outperforms other more sophisticated devices when used for creation of non-photorealistic rendering (NPR) images. The author compares the sensitivity of a wired two-button mechanical roller ball mouse versus a high precision laser mouse and profiles the futility of the wireless self-powered optical mouse, as well as the pen tablet, for developing NPR images.

Fig. 8. Atman Victor, Brainstorm

7. Learning, Education

Examples presenting the integrative curricula for artists, designers, computer scientists, and engineers, and the interdisciplinary projects for software engineering students illustrate the growing belief that artists and engineers reinforce each other by complementary strengths.

Vinod Srinivasan, Donald House, Mary Saslow, and Carol LaFayette (2007, pp. 839-844) describe basic training for digital artists in the Texas A&M Visualization Program. At Texas A&M University, a Master of Science program intermixes concepts, students, and faculty from art, computer science, architecture, and engineering. Artists, designers, computer scientists, and engineers are sent through the same curriculum. Students are later hired as ‘technical directors’ and ‘technical artists’ in the special effects, computer animation, and electronic games industries. Art students take ‘Computing for Visualization’ I (C programming, system utilities, shell scripting, and introductory OpenGL), and II (C++ programming, vector and matrix algebra, mathematics of spline curves, and intermediate OpenGL). Computer Science students take ‘Concepts for Visual Communication I & II, described in detail). Best practices for both kinds of courses are discussed.

Letizia Jaccheri and Guttorm Sindre (2007, pp. 925-934) describe an interdisciplinary program for software engineering students at the Norwegian University of Science and Technology, Trondheim. The program includes three interdisciplinary team projects: performing the design, coding, and testing the teacher-supplied task (a software engineering assignment); performing a problem and requirements analysis, design and a partial implementation of the system (a system development project); and focusing on interdisciplinary work (the Experts in Team project). Students learn to approach software as not a goal in itself and cooperate with others with different expertise.

Martin Constable (2007, pp. 850-859) analyzes a digital image in a way that is useful to a student of art. Visual feedback in digital practice, equivalent to tricks that are used in traditional painting, support artists in evaluating and judging the success in their art. In digital painting, tools that are used to change, enhance, or distort an image can be also used to reveal its visual structure and thus enhance an understanding of its strengths and shortcomings. These techniques include: gaining a fresh look by applying ‘flip horizontal’ command or zooming tool in Photoshop; considering tone and color separately through de-saturating and posterizing the image or separating the tone into different passes; temporarily changing color mode; extracting linear elements of the image; reviewing layers to examine space in an image; and otherwise mapping aspects of an image.

8. Summary and Conclusion

Topics selected for this review reflect only a few selected themes related to the whole field of visual analytics, visualization, information visualization, and visual data mining. Also, they do not cover the broad spectrum of the 11th International Conference on Information Visualization in Zurich, Switzerland. Specialized sessions held at the Conference included: visual analytics, themes related to information visualization: theory, techniques, usability, applications, collaborative visualization, web visualization, large-scale visualization, visualization in built and rural environments, information visualization in biomedical informatics, and design visualization, and themes related to knowledge visualization: theory, new classifications, applications, indigenous knowledge visualization, and knowledge domain visualization. Other sections related to geo-visualization visual data mining, HCI, interaction design for information visualization, applications of graph theory, augmented, mixed, and virtual reality, multimedia and e-learning. Proceedings of the 11th International Conference on Information Visualization, edited by E. Banissi, R. A. Burkhard, G. Grinstein, U. Cvek, M. Trutschi, L. Stuart, T. G. Wyeld, G. Andrienko, J. Dykes, M. Jern, D. Groth, A. Ursyn, A. Faiola, A. J. Cowell, and M. Hou, have been published by IEEE Computer Society Press.

References

Andrews, K., Putz,W., & Nussbaumer, A. (2007). The Hierarchical Visualization System (HVS). In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 257-262). IEEE Computer Society Press.

Bates-Brkljac, N. (2007). Investigating perceptual responses and shared understanding of architectural design ideas when communicated through different forms of visual representations. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 348-353). IEEE Computer Society Press.

Benoy, F., & Rodgers P. (2007). Evaluating the Comprehension of Euler Diagrams. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 771-778). IEEE Computer Society Press.

Bertschi, S. (2007). Without Knowledge Visualization? Proposing a Deconstructivist Approach to Metaphor, Meaning and Perception. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 342-347). IEEE Computer Society Press.

Bertschi, S. & Bubenhofer, N. (2005). Linguistic Learning: A New Conceptual Focus in Knowledge Visualization. In 9th International Conference on Information Visualization, London (pp. 383-389). IEEE Computer Society Press. 

 

Burkhard, R. A. (2005). Knowledge Visualization: The use of Complementary Visual Representations for the Transfer of Knowledge – A Model, a Framework, and Four New Approaches. D.Sc. thesis. Swiss Federal Institute of Technology (ETH Zurich).

Cawthon, N. & Wande Moere, A. (2007). The Effect of Aesthetic on the Usability of Data Visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 637-643). IEEE Computer Society Press.

Durupinar, F., & Güdükbay, U. (2007). A Virtual Garment Design and Simulation System. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 862-867). IEEE Computer Society Press.

Constable, M. (2007). Analyzing a Digital Image in a Way that Is Useful to a Student of Art. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 850-859). IEEE Computer Society Press.

Eppler, M. J., & Platts, K. (2007). An Empirical Classification of visual methods for Management: Results of Picture Sorting Experiments with Managers and Students. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 335-341). IEEE Computer Society Press.

Erbacher, R. F. (2007). Exemplifying the Inter-Disciplinary Nature of Visualization Research. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 623-630). IEEE Computer Society Press.

Fathulla, K., & Basden, A. (2007). What Is a Diagram? In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 951-956). IEEE Computer Society Press.

Hamdi, M. S. (2007). Semantic Map Based Web Search Result Visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 222-227). IEEE Computer Society Press.

Hook, P. A. (2007). Domain Maps: Purposes, History, Parallels with Cartography, and Applications. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 442-446). IEEE Computer Society Press.

Hsin Hsin, L. (2007). A Non-photorealistic Rendering Images by a Handheld Device. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 845-849). IEEE Computer Society Press.

Jaccheri, L., & Sindre, G. (2007). Software Engineering Students Meet Interdisciplinary Project Work and Art. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 925-934). IEEE Computer Society Press.

Jamieson, R. & Alexandrov, V. (2007). A Data Forest: Multi-Dimensional Visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 293-298). IEEE Computer Society Press.

Jern, M. (2007). An Information Visualization Approach to Retail Space Management (VisMT). In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 109-116). IEEE Computer Society Press.

Jern, M., & Franzen, J. (2007). Integrating InfoViz and GeoViz Components. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 511-518). IEEE Computer Society Press.

Kienreich, W., Zechner, M., & and Sabol, V. (2007). Comprehensive Astronomical Visualization for a Multimedia Encyclopedia. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 363-368). IEEE Computer Society Press.

Kosara, R. (2007). Visualization Criticism – the Missing Link between Information Visualization and Art. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 631-636). IEEE Computer Society Press.

Lau, A., & Vande Moere, A. (2007) Towards a Model of Information Aesthetics in Information Visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 87-92). IEEE Computer Society Press.

Lengler, R. (2007). How to Induce the Beholder to Persuade Himself: Learning from Advertising Research for Information Visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 382-392). IEEE Computer Society Press.

Lin, X., Bui, Y., & Zhang, D. (2007). Visualization of Knowledge Structures. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 476-481). IEEE Computer Society Press.

Peters, G. (2007). Aesthetic Primitives of Images for Visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 316-325). IEEE Computer Society Press.

Rafi, A., Tinauli, M., & Izani, M. (2007). High Dynamic Range Images: Evolution, Applications, and Suggested Processes. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 877-882). IEEE Computer Society Press.

Rakkolainen, I. (2007). How Feasible Are Star WarsTM Mid-air Displays? In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 935-942). IEEE Computer Society Press.

Roberts, J. C. (2007). State of the Art: Coordinated and Multiple Views in Exploratory Visualization. In International Conference on Coordinated and Multiple Views in Exploratory Visualization, associated and collocated with International Conference on Information Visualization (iV07), (pp. 61-71). IEEE Computer Society Press.

Schulz, H-J., & Schumann, H. (2006). Visualizing Graphs - A Generalized View. In 10th International Conference on Information Visualization, London, England, (pp. 166-173), IEEE Computer Society Press.

Shneiderman, B. (1996). The eyes have it: A task by data type taxonomy of information visualizations. In Proceedings of IEEE Visual Languages ’96, (pp. 336-343), IEEE Computer Society Press.

Silva, S., Madeira, J., & Sousa Santos, B. (2007). There Is More to Color Scales than Meets the Eye: A Review on the Use of Color in Visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 943-950). IEEE Computer Society Press.

Slingsby, A., Dykes, J., Wood, J., & Clarke, K. (2007). Interactive Tag Maps and Tag Clouds for the Multiscale Exploration of Large Spatio-temporal Datasets. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 497-504). IEEE Computer Society Press.

Spoerri, A. (2007a). Visual Mashup of Text and Media Search Results. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 216-221). IEEE Computer Society Press.

Spoerri, A. (2007b). Coordinating Linear and 2D Displays to Support Exploratory Search. In International Conference on Coordinated and Multiple Views in Exploratory Visualization, associated and collocated with International Conference on Information Visualization (iV07) (pp. 16-26). IEEE Computer Society Press.

Srinivasan, V., House, D., Saslow, M., & LaFayette, C. (2007). Basic Training for Digital Artists in the Texas A&M Visualization Program. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 839-844). IEEE Computer Society Press.

Suvanaphen, E., & Roberts, J. C. (2007). Visualizing Evolving Searches with EvoBerry. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 238-244). IEEE Computer Society Press.

Vernik, R., Von Itzstein, G. S., & Bouchard, A. (2007). Imago: An integrated prototyping, evaluation and transitioning environment for information visualization. In 11th International Conference on Information Visualization, Zurich, Switzerland pp. 17-22). IEEE Computer Society Press.

Wang, X., Chen, R., Gong, Y., & Hsien, Y-T. (2007). Experimental Study on Augmented Reality Potentials in Urban Design. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 567-572). IEEE Computer Society Press.

Weber, W. (2007). Text Visualization – What Colors Tell About a Text. In 11th International Conference on Information Visualization, Zurich, Switzerland (pp. 354-359). IEEE Computer Society Press.

[1] This review is focused on papers presented at the 11th International Conference on Information Visualization, Zurich, Switzerland, 4-6 July 2007, and the Proceedings published by the IEEE Computer Society. The Symposium and Gallery of Digital Art (D-Art) accompanies the Conference: art papers and artwork can be seen at: http://www.graphicslink.co.uk/DART/IV&CGIV07Gallery/index.html).

 
 
 
About | Issues
© NMEDIAC & individual NMEDIAC authors, editors, and programmers
.
home issues