The Community for Technology Leaders
RSS Icon
Subscribe
Issue No.01 - January/February (2012 vol.32)
pp: 20-21
Published by the IEEE Computer Society
Giuseppe Di Battista , Third University of Rome
Huamin Qu , Hong Kong University of Science and Technology
ABSTRACT
This article introduces the articles in the special issue on visualization applications and design studies. The articles cover both traditional application domains such as visualization of scientific simulations and emerging ones such as visualization of text streams, image collections, trajectories, and political phenomena.
Applications have been driving visualization research from the early days on. Researchers have developed countless impressive techniques to visualize and interact with large, complex data from areas as diverse as scientific simulation, medicine, and sociology. Application domains provide data and problems for visualization research and the battlefields to assess new techniques' effectiveness. Progress in visualization research has also led to a wealth of applications crossing multiple disciplines. Researchers are collecting new types of data, new problems are emerging, and domain experts are more often employing visualization to gain insight into these data types and solve their problems.
This special issue contains five articles covering both traditional application domains such as visualization of scientific simulations and emerging ones such as visualization of text streams, image collections, trajectories, and political phenomena.
Scientists have long used visualization to communicate results from simulations such as in computational fluid dynamics. However, simulations are becoming increasingly complex, posing special challenges in visualization system design. In "Visual Analysis of Particle Behaviors to Understand Combustion Simulations," Jishang Wei and his colleagues present a visualization system that can help scientists better understand the turbulent dynamics in combustion processes. The system tackles a challenging visualization problem: the simulations involve millions of moving particles, and each particle changes its position and thermochemical state. The authors present a dual-space method to visualize particles' movement in the physical space and their attribute evolution in the phase space. To address the large-data problem, a semisupervised-learning technique classifies the attribute evolution curves into distinct groups, each of which can then be visualized without clutter. This article provides valuable lessons for designing similar systems in the future.
Most information is stored as text, and streaming text data such as email and news has become common. In "Real-Time Visualization of Streaming Text with a Force-Based Dynamic System," Jamal Alsakran and his colleagues introduce Streamit, an interactive visualization system for monitoring and analyzing text streams such as news from the New York Times. Streamit uses a force-directed simulation to reveal document clusters and out-liers, dynamically inserting each newly arrived document into the display. The system has some nice features such as adaptive keyword vectors to reflect content changes in the incoming documents, adjustable keyword importance to indicate a keyword's significance at a certain time, topic modeling to reveal documents' higher-level semantic meanings, and GPU acceleration for fast simulation and immediate response. These features make Streamit a powerful interface for in-depth analysis of streaming text data.
Exploring image collections is notoriously difficult, especially for nonexperts. Most interfaces rely on a metadata-based search that can be difficult to understand and does not allow quick comprehension of the topics covered. In "ImageHive: Inter-active Content-Aware Image Summarization," Li Tan and his colleagues describe a technique that uses a creative mixture of image segmentation, layout by similarity, and interaction to generate visual summaries of image collections. The summaries look nice, preserve important relationships between images, and present in a glimpse a fair summary of a collection. We hope the technique becomes as popular for image collections as Wordle has become for text collections.
Exploring the trajectory of moving objects, such as in vessel traffic or animal migration, is challenging. It requires aggregating many trajectories to make sense of trends, as well as investigating outliers and special events. In "Interactive Density Maps for Moving Objects," Roeland Scheepens and his colleagues introduce an inventive technique that divides the visualization into three steps: filtering, computing the density field, and rendering and compositing. This pipeline offers an excellent trade-off between expressive power to effectively analyze trajectories and conceptual simplicity.
Finally, in "Two Visualization Tools for Analyzing Agent-Based Simulations in Political Science," R. Jordan Crouser and his colleagues present two original systems, designed in collaboration with domain experts, to support inquiry and inference by social scientists using agent-based simulations to model political phenomena. The systems can help scientists better understand the forces at work in social and political systems, which can in turn enable them to better inform decision-makers and international policy. Each system uses a coordinated multiview architecture, letting analysts customize the views to suit their analytical process.
All these articles attest how important visualization has become to solving real-world applications. We were impressed by the depth and breadth of the articles we received for this special issue. Our field is growing and maturing; we hope you enjoy reading the articles as much as we did.
Giuseppe Di Battista is a professor of computer science in the Third University of Rome's Department of Computer Science and Automation. His research interests include computer networks, graph drawing, and information visualization. Di Battista has a PhD in computer science from Sapienza University of Rome. Contact him at gdb@dia.uniroma3.it.
Jean-Daniel Fekete is the scientific leader of the INRIA AVIZ (Analysis and Visualization) research team. His main research areas are visual analytics, information visualization, and human-computer interaction. Fekete has a PhD in computer science from Université Paris-Sud. He's a member of ACM and IEEE. Contact him at jean-daniel.fekete@inria.fr.
Huamin Qu is an associate professor in the Hong Kong University of Science and Technology's Department of Computer Science and Engineering. His main research interests are visualization and computer graphics. Qu has a PhD in computer science from Stony Brook University. Contact him at huamin@cs.ust.hk.
680 ms
(Ver 2.0)

Marketing Automation Platform Marketing Automation Tool