detail of a section of the Wikipedia map

Who can possibly grasp the contents of Wikipedia? In the six years since it was launched, the "free encyclopedia of everything" has grown to include 2.1 million articles in English alone, written by more than 150,000 contributors, with a combined 7.5 million articles in Wikipedias of all languages.

But Katy Börner knows what Wikipedia looks like. Borner, the Victor H. Yngve Professor in the School of Library and Information Science at Indiana University Bloomington, is also director of the Information Visualization (InfoVis) Laboratory at IUB. With her InfoVis colleagues Bruce Herr and Todd Holloway, Borner has analyzed and "visualized" the network of knowledge that is Wikipedia in English (http://en.wikipedia.org). They categorized Wikipedia articles, then gathered images from the most popular pages (approximately one picture for every 64 articles). The researchers laid out the article images on a circular grid, positioning similar, linked articles close together. The result is a mosaic revealing Wikipedia's most popular knowledge areas.

Colored dots on top of the grid indicate how often and how recently the articles were edited. Large dark dots indicate lots of activity. Among the top 10 most actively revised Wikipedia articles are about Jesus (#1), Adolf Hitler (#2), and Britney Spears.

The Wikipedia picture is fascinating and fun. It's also potentially quite useful to Wikipedia administrators, according to collaborator Bruce Herr. In a May report on the mosaic in New Scientist, Herr suggested that updating the image in real time could allow Wikipedia's administrators to track the articles being most actively edited and thus spot controversial arguments that may be popping up.

But the methods of data analysis and visualization used by Börner and her colleagues are widely relevant beyond Wikipedia, as Borner pointed out in a recent paper presented at the 2007 Information Visualization International Conference in Switzerland. Imagine a research group or center or laboratory made up of 50 people, write Borner and her IU colleague Thomas Neirynck. Then consider all the projects, courses, publications, presentations, events, datasets, software, hardware, and funding sources associated with such a group. "To our knowledge," say Borner and Neirynck, "there exists no system designed to support the management of scholarly data relevant to the operation of a research unit and the communication of this data to different stakeholders such as group leads and members, information technology staff, funding agencies, or other scholars."

Using her own InfoVis Lab (http://cns-nd3.slis.indiana.edu/ivl/) as a guinea pig, Börner has implemented a system of data representation, analysis, and visualization tailored to the needs of various research stakeholders. After data about scholarly activities such as grants, teaching, publications, and research has been collected, tagged, and stored, the data can then be displayed graphically in tables, graphs, network diagrams, or geospatial maps. These different kinds of visualization expand access to and dissemination of scholarly and scientific information.

For example, one network diagram for the InfoVis lab shows all collaborations on papers produced by staff and students over a year's time, revealing linkages between research groups that may otherwise go unnoticed. A network display of funding awards over five years simultaneously reveals not only who has received awards, but the amount of the awards, the year received, and who is working on which projects together.

The Wikipedia mosaic and the InfoVis Lab data visualization project are just two pieces of a much larger intellectual mission that Börner and her colleagues are pursuing-a mission that is nothing less than to visualize all of humankind's scholarly knowledge and expertise in digital form. In other words, to make sense of everything we know.

As Börner and her colleagues put it in a 2007 "sketch" on mapping science, "Today's number of researchers exceeds the number of researchers ever alive before. Some areas of science produce more than 40,000 papers a month. No human or machine can process and make sense of this enormous stream of data, information, knowledge, and expertise. ... Maps of our planet have guided humankind's explorations for centuries. Maps of science can guide our scholarly endeavors ... At the global level, science maps provide a birds-eye, holistic view of how all areas of science are interrelated."