Symposium on the Future of Visualization
William Ribarsky, Director, Charlotte Visualization CenterComplete Symposium Schedule
On May 1 and 2, 2006, we held the Symposium on the Future of Visualization at UNC Charlotte. The Symposium was in celebration of the grand opening of the Charlotte Visualization Center. It was thus an opportunity to look ahead as we stepped across the threshold into a new venture of regional importance that we hoped would eventually achieve national impact. We successfully seized this opportunity with an event attended by 300 registrants, having a list of visionary speakers unlikely to be heard together anywhere else, and producing a wide, exciting spectrum of ideas to be discussed and thought about.
The Web pages assembled here provide a multimedia record of this event so that a wide audience can partake of its gathered insights. This essay is both a brief guide and a personal rumination on what transpired and where it fits in current trends in visualization.
This Symposium was just one of several recent attempts to assess where visualization stood now and where it should go from here. Visualization is a young field (less than 20 years old if one takes the seminal 1987 NSF report, "Visualization in Scientific Computing", as the starting point). It is reasonable that there should be periodic reassessments of the whole field, and that requirement is what is behind the "NIH/NSF Visualization Research Challenges" report now available. However, beyond periodic assessment, new needs and challenges have been found for visualization, as demonstrated by the recent book, "Illuminating the Path: The Research and Development Agenda for Visual Analytics." This agenda was developed by a group of leading researchers in visualization, statistics, and other fields in collaboration with colleagues from government and industry. The agenda was backed by the U.S. Department of Homeland Security in response to pressing needs in intelligence analysis, emergency response and disaster prevention, and border security. However, it is clear that the agenda and the visual analytics field that it created stretch well beyond the needs of homeland security. Further, it is evident that though significant achievements could be expected in the first years pursuing the agenda, it would take much longer and much new research and intellectual activity to achieve the deeper goals of the agenda, such as the intimate collaboration of human and computer on complex exploratory analysis problems.
What are the substantial research challenges and needs that now confront visualization? In some ways, they are extensions of the original problems when the field was first created, having to do with the handling of large data and the control of simulations, models, or instruments that produce that output. But now the production of data has become quite overwhelming and complex, and new sources such as the Web have made knowledge creation and sharing ubiquitous. The result is that a class of problems have emerged that are intractable using traditional means but for which interactive visualization shows much promise. These problems are exploratory in nature and require timely assessments and decisions. Humans must thus be intimately involved in the data collection, hypothesizing, and knowledge creation; fully automated techniques will not succeed. Yet individual humans do not have the capacity to gain both overviews and, where necessary, intimate understanding of all they must assess, since they must typically discover what's important and do not know this beforehand. Nor, in many cases, are there time and resources to parallelize these analysis problems by applying large groups of humans to them. (Which is not to say that collaboration among smaller groups of people and across traditional administrative or disciplinary boundaries is not, in many cases, essential.) Intelligence or business analyses that may require a user to watch over or assess a large segment of open sources, such as the Web or broadcast video, are such problems. A wide range of other problems are of this type, such as tasks in bioinformatics, where users need to discover the details of gene or protein function by applying successive, iterative analyses and comparing them to a vast array of genomic annotations and previous results. Highly interactive and general tools, such as IN-SPIRE for text exploration and analysis, have had success with these types of problems. Successes that have led to the visual analytics program, whose research thrusts will go well beyond what these tools provide.
Overview of the Symposium
The Symposium showed that visualization is a wide, growing, and vibrant field. Invited speakers explored topics such as visual thinking, medical diagnosis applications, the promise and challenges in the new field of visual analytics, the prospects for automating graphic design, the idea that basic perceptual principles can be identified and applied to greatly improve visualization of large scale data, fast, precise, and more effective diagnostic techniques through medical visualization, advancing beyond data or information visualization to knowledge visualization, and the idea that careful rendering can be used to explore and understand the creative impulses and goals of someone who designs in space (in this case an architect) even when those designs have never been built. These talks cover just a part of what the field of visualization contains, yet they address some of its most pressing questions. They also show again that the field of visualization is not just for computer scientists and that, in fact, it reaches its deepest insights and most effective uses when it draws from physical scientists, psychologist, artists, engineers, statisticians, architects, and others. The breadth on display also illuminates some of the strains and challenges visualization faces. A main question is, if visualization contains significant contributions from other fields, how do we define it? What are its unique principles and practices? This is the flip side of the oft-stated idea that visualization provides a "common picture", bringing together investigators from different disciplines and fusing data from multiple sources. It is perhaps because of these strains that visualization is still not very well-defined as a discipline. There are no widely used textbooks, for example (although information visualization is farther along in this regard than the rest of the field). Many of us in the field think it is past time to fill this gap and that we must start determining and agreeing upon the principles and practice of the field. Only then will the full flowering of the field occur. Visual analytics is trying to squarely face the issues of courses, curricula, and developing a research agenda, but more must be done.
Go to page: 1 2 3 Next
|
Symposium Online Content
|
||||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
Presentations by Some Future Leaders
|
||||||||
|---|---|---|---|---|---|---|---|---|
|
