Skip to main content


Issue Browser: An Interactive Search Environment for Domain-Specific Knowledge Acquisition via Multimedia Data


IssueBrowser is an interactive visualization system that aims to enhance the knowledge acquisition capabilities of its users with regard to a particular knowledge domain: the 2008 Presidential Election. Since the task of distilling large amounts of information about an issue, such as the 2008 Presidential Campaign can represent a cognitive challenge for most users, IssueBrowser seeks to simplify the process of gathering and filtering through automation. IssueBrowser was designed to run as a public installation on a university campus. Our goal is to facilitate critical engagement with political issues through a multimedia collection of loosely organized data. We hope to inspire users, and especially young first-time voters, to see issues from a variety of perspectives and to inform their opinions about those issues.

Drawing on Internet data sources as diverse as candidate campaign sites, YouTube, Yahoo! News, think tanks and Wikipedia, IssueBrowser sorts information by keyword and presents it in bite-size chunks that users can then use to form an understanding of the campaign. By performing this task for the user, IssueBrowser contributes to a reduction in cognitive load. Unlike a traditional news website, IssueBrowser reduces the amount of extraneous information presented to the user, allowing the user to devote more attention to relevant materials.

IssueBrowser features a modular architecture that gathers data related to issues and/or candidates from the sources mentioned, filters out irrelevant portions, and adds meta-information so that the data is usefully connected together. We present an innovative visualization system that allows users to interpret data (and connections between them) from various perspectives and from various sources perspectives and sources to which users might not be exposed during a normal web search. Our layout algorithm has both temporal and spatial aspects that work together to balance simultaneously both a broad view of a range of issues as well as increasingly detailed information about particular aspects of an issue or a cluster of issues. In addition to displaying text (articles, quotes, facts) and images (video, photos), we provide a novel interaction system which allows multiple users to choose collaboratively the various aspects of particular issues they are interested in exploring. Users influence the flow of new information by “pinning” particular data— that is, forcing new queries to return data to be related to the currently selected data.

Our interdisciplinary team includes students from Computer Science, Geography, Electrical and Computer Engineering, Education, and Media Arts and Technology. Together, we identified common research goals and decided to create a large scale, interactive public installation for our University library. Facets of our ambitious project simultaneously represent expertise from our varied disciplines. For example, informed by our Education student’s research that multiple information options and sources can be cognitively overwhelming, we set about to create a system that simplifies the information gathering process. To do so, we drew upon the expertise of our Geography and Computer Science students to develop a database that organizes and analyzes information harvested from various online sources, including, but not limited to candidate websites, YouTube, Yahoo! News, think tanks and Wikipedia. Text mining algorithms so far include quote extraction and attribution, and a similarity measurement that classifies any document (or aggregation of documents) as being more or less about each of 34 campaign issues. The latter allows comparisons of political speech across candidate, party, region and time period.. To complement our text mining efforts, our Electrical and Computer Engineering student and a Computer Science colleague developed a search to identify key frames in our video data and extract relevant segments to be used as part of our visualization. Our Computer Science and Media Arts and Technology students have created a xx graph to visualize the information and developed an engaging Wii-enabled interactive interface for our public installation.

While each component could certainly be created in a disciplinary vacuum, their combination in a single system is made possible through our interdisciplinary collaboration. By sharing disciplinary perspectives, we applied our expertise, but also modified our approaches to complement our team’s efforts. We often received questions about our approaches that would not typically be asked in our own discipline. In the end, we created a novel search tool that builds upon a user’s previous searches to deliver candidate information with increased relevance and specificity. Our goal was to help college-level voters become more informed about the candidates and their issues. Through a combination of algorithms, designs, user studies, and interactive visualizations, we created a public installation that delivers this information in a seemingly simple way.

Address Goals

NSF-funded researchers at the University of California, Santa Barbara are developing a new information visualization program called IssueBrowser to encourage investigation of current news topics in a public space. The IssueBrowser visualization is displayed at the east entrance to the Davidson Library on the University of California, Santa Barbara campus. Students passing by can stop in front of the display wall and, using a wireless pointer, select people and issues which interest them. The system provides relevant text, images, and video culled from the Internet about the selected topics so that a user can delve deeper into a wide variety of source materials.

The visualization techniques developed for IssueBrowser can be applied to any knowledge domain. Currently we are focusing on news of the American presidential election campaign. However, we can easily direct our information gathering methods to science news websites. Once we have collected enough source material, we use automatic machine learning techniques to organize the material into individual topics, which can then be browsed hierarchically through IssueBrowser. Our public installation space in the campus library provides a unique opportunity for the NSF strategic goal of science learning among the general public.

Our system offers an alternative to web interfaces by displaying a broad spectrum of documents at the same time. Unlike a web interface, where one must specifically navigate to source websites and view pages one at a time, IssueBrowser automatically presents several sources at once, spread across a large display wall. This allows for comparison between sources, and may present sources that the user would not normally find in their web search

Currently, we have directed our data collection algorithms to focus on the current American presidential election campaign. However, past versions of our system focused on world news, from all countries across the globe. By visually connecting headlines and images to their source locations on a world map, we aimed to increase awareness of global issues with our visualization. This version of IssueBrowser (then entitled “Spheres of Influence”) was presented at the prestigious Association of Computing Machinery Special Interest Group on Graphics (ACM SIGGRAPH) digital art exhibit in San Diego. There, several thousand attendees interacted with the visualization and learned about world news through our unique system.