It is an honor and a pleasure to present in this theme issue extended versions of the best papers from the recent Visual Analytics Science and Technology Symposium (VAST 07), held in conjunction with IEEE Visualization 2007 (October 28–November 1, 2007 in Sacramento, CA). Now in its second year, VAST is the premier international conference in visual analytics. VAST 07 had 24 papers accepted, presented, and published in the proceedings. The number and quality of papers is noteworthy since there has been increased competition from journal special issues devoted to visual analytics and several conferences where visual analytics papers are encouraged. The overall picture is of a strong and growing field. The seven papers selected span important visual analytics topics such as investigative analysis, emergency response, collaboration, text and media analyses, social analytics, and interaction. These are not only rich research topics, they also involve techniques of importance to those who need to apply visual analytics tools and methods. Several of these papers and also other papers from VAST 07 reflect growing collaborations between visualization scientists and those who do analytics, either in mathematical and statistical methods, social analyses, financial analytics, or other areas. We expect that these collaborations will grow and in the future there will be leadership from the analytics side. Based on ongoing research in the visual analytics community and interests expressed in discussions at workshops and other venues, in the future we also expect to see more papers in the important areas of analytical reasoning, cognitive models, the fundamental properties of interaction, and presentation visual analytics for decision makers and others.
In 2007 the VAST Area Chairs acted as primary reviewers, handling the review process for all the VAST submissions and ultimately selecting the seven most outstanding papers for inclusion in this IVS theme issue. Each author was then required to add substantial new material beyond the VAST paper. This material included new examples and results, extensions of previous methods, evaluations, or explanations of impacts of the work. The papers were then reviewed to make sure that they met the criteria on new material and appropriateness for IVS. We thank the many reviewers who provided objective and useful reviews throughout the process.
The first paper in this issue, 'Stories in GeoTime', by Thomas Kapler, Ryan Eccles, Robert Harper, and William Wright, was awarded the Best Paper at VAST by a panel of visual analytics experts. The paper reports on a system that demonstrates geotemporal patterns in stories so that intelligence analysts can conceptualize threats and understand patterns in support of their reasoning process. The powerful capabilities in this paper can be used generally in investigative analysis.
The next paper, 'DataMeadow: a visual canvas for analysis of large-scale multivariate data', by Niklas Elmqvist, John Stasko, and Phillipas Tsigas studies issues in the visual analytics of multiple large-scale multidimensional data sets. The paper describes DataMeadow, a way to interactively construct visual queries using graphical set representations called DataRoses. Importantly, the authors' method provides an annotatable visual history of the analysis to facilitate dissemination of the results. An evaluation of the system is provided, which uses a qualitative expert review involving two visualization researchers.
The third paper is 'Visual analysis of controversy in user-generated encyclopedias' by Ulrik Brandes and Jurgen Lerner. They attack the multiplicity of viewpoints offered in the Wikipedia collaborative authoring environment. Their visual analytics approach reveals rich insights into controversies, including who are the dominant authors, what roles they play, and how they interact. As collaborative authoring environments become more common, such analytical tools will be important.
The next paper is 'Design considerations for collaborative visual analytics' by Jeffrey Heer and Maneesh Agrawala. The authors note that though much research to date in visual analytics assumes a single-user focus on perceptual and cognitive processes, in practice, the sensemaking process is often a social one and should use approaches that support social interaction. The authors suggest that appropriate collaboration mechanisms are not well understood and they develop considerations for guiding the design and evaluation of collaborative visualization systems.
The next paper is 'Scalable and interactive visual analysis of financial wire transactions for fraud detection' by Remco Chang, Alvin Lee, Mohammad Ghoniem, Robert Kosara, William Ribarsky, Jing Yang, Evan Suma, Caroline Ziemkiewicz, Daniel Kern, and Agus Sudjianto. This paper shows how visual analytics can be applied to the very difficult activity of detecting suspicious activity and possible fraud in the wire transfers at a very large bank. Hundreds of thousands or more transactions must be monitored over a given period of time and, without the visual analytics tools that were developed, it is difficult to separate suspicious activity from regular activity. This paper demonstrates what a team of bank experts and visualization researchers can accomplish.
The sixth paper, 'Mobile analytics for emergency response and training', by Sung Ye Kim, Ross Maciejewski, Karl Ostmo, Edward Delp, Timothy Collins, and David Ebert describes a mobile visual analytics tool to help enhance situational awareness and support rapid decision making in the context of emergency response events. Their system generates 2D/3D visualizations on hand-held devices such as a PDA to show personnel-related information, situational, and static scene-related information. It also includes playback functionality for camera-outfitted personnel, along with fast-forward/rewind capabilities for reviewing events. Their system was used in the evaluation of two scenarios, including a simulated evacuation of a nightclub during a fire.
The last paper is 'Visual analytics approach to user-controlled evacuation scheduling' by Gennady Andrienko, Natalia Andrienko, and Ulrich Bartling. The authors apply visual analytics to time-critical decision making that has a geotemporal aspect. The analysis includes automatic scheduling based on a genetic algorithm, which is integrated with a visualization of transportation schedules. This paper shows how to attack a class of complex, urgent decision-making problems that will be essential to solve.
The field of visual analytics is growing in scope and depth. In addition to the United States, more research is being reported from Canada, Europe, and now Asia. We encourage the readers of this issue to become involved. Good opportunities for the publication of research results will continue to abound, such as VAST 08, which will be held in conjunction with VisWeek 2008 in Columbus, OH. (VisWeek is the new name for the combined IEEE Visualization and InfoVis Conferences and VAST.)
