Symposium on the Future of Visualization
William Ribarsky, Director, Charlotte Visualization CenterComplete Symposium Schedule
Continued from page 1
One clear trend is that visualization is addressing more complex tasks. Rather than just revealing the content and correlations in a dataset, certain visualizations are now taking on a whole investigative, exploratory, or diagnostic process. An outstanding example of this is Arie Kaufman's Virtual Colonoscopy project that must consider the whole process from data organization, 3D reconstruction, and analysis for visualization through proper presentation via a subtly interactive visual interface to diagnostic investigation. All these components were the product of sustained basic research in interactive volume visualization guided by a clear goal and long-term collaboration with medical professionals. The resulting visualization and diagnosis system is a uniquely capable tool (providing, for example, the capability to identify, measure, and look at the internal structure of any tumor or polyp); it is more immediate, less difficult for both patient and doctor than other methods, and it saves lives. Moreover, it is leading to similar 3D reconstruction and diagnostic tools for other organs and for blood vessels.
The trend towards more complex investigative and exploratory tasks has led to a fresh look at underlying visualization principles. For these tasks, it becomes clear that one cannot just consider the basic perceptual principles underlying interactive rendering or improved methods for constructing and rendering graphical models of the data. One must consider the whole perceptual/cognitive spectrum taking into account reasoning, hypothesis building, and knowledge creation processes. The problems mentioned above, where the user must be intimately involved in the analysis process, are tasks of this type. Indeed, open ended problems beyond a certain size and complexity are of this type, ranging from scientific research problems to investigative reporting.
Three of the Symposium talks addressed these issues in interrelated ways. Stu Card talked about Using Vision to Think and in particular about the Sensemaking approach. Sensemaking is a framework developed over the past several years for modeling and understanding visual analyses that lead to decisions, starting with evidence gathering, leading to hypothesis building, and ending with presentation for action. It is thus both a working approach for exploratory analysis and a possible basis for a broader theory of analytical reasoning and human-information discourse. Card has applied this framework to intelligence analysts since they are the "jet pilots of sensemaking". This investigation has led to an enriched understanding of the components of the sensemaking structure, which turn out to be linked to each other in a highly iterative set of loops. The investigation also indicates leverage points where further research in the visual analysis process could lead to significantly improved results. Others are now building on this model to develop more detailed analytical reasoning capabilities. Bill Ribarsky argues for the development of a Knowledge Visualization approach that goes beyond data or information visualization. By dealing with knowledge artifacts rather than data artifacts, one is dealing more directly with the raw material of reasoning and understanding. One can more precisely assess the value of visualizations and may be able to create visualizations of much higher value. Further, the approach gives a clearer idea of the meaning and use of interactivity in exploratory analysis and knowledge creation. As researchers continue to develop this approach, a set of design principles for interactive knowledge visualization should emerge. Jim Thomas addressed the research agenda of Visual Analytics directly, showing that successful pursuit of the agenda will open the way to solving grand challenge problems in research and applications. The visual analytics program has rapidly grown into a significant source of intellectual activity and research funding for the visualization community. It is the catalyst for new activities at agencies such as NSF and at companies. New venues for publishing research results are being created, and a new conference (the IEEE VAST Symposium) has been established. A challenge for the visualization community will be to decide where visual analytics fits, since it is in one sense a component of the larger visualization field and in another sense broader than visualization because it brings together other disciplines (such as statistics and cognitive science). One challenge for the visualization community will be to determine the extent of the visual analytics curriculum and distinguishing it from the visualization curriculum.
Some of the speakers took fresh looks at longstanding problems in visualization. Pat Hanrahan discussed Automating Basic Graphics Design. This is a longstanding issue because most design for visualization is still done by hand. This is in spite of efforts to develop modular visualization environments such as IBM Explorer, AVS, or Iris Explorer, or toolkits such as VTK. Main goals of these approaches have been to make high quality visualization algorithms widely available and to permit non-experts to create visualizations tailored to their specific purposes and applications. However, these goals have turned out to be hard to achieve in practice. Once users get away from basic visualizations, these approaches become significantly harder to enact, and the plethora of modules available can be overwhelming. Furthermore, the approaches do not help much in designing or laying out good, effective visualizations. Yet, as Pat points out, visualizations and the need for good visualizations are becoming ubiquitous. Under these circumstances, an automated approach to task-specific design optimization will be extremely useful. It will also open the door to more and better device-dependent graphics, so that visualization applications that run on your desktop can also run successfully on extreme devices such PDAs or wearable computers. Finally, some basic principles for graphics design are emerging and will likely continue to emerge from this approach.
Go to page: Previous 1 2 3 Next
|
Symposium Online Content
|
||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
Presentations by Some Future Leaders
|
||||||||
|---|---|---|---|---|---|---|---|---|
|
