They're not your typical cartographers.  Researchers from the National Research Council's Canada Institute for Scientific and Technical Information (NRC-CISTI) and Carleton University have produced a map of science that aims to make it easier to navigate through millions of scientific research articles.

Created using NRC-CISTI's large collection of digital scientific, technical and medical (STM) articles, this map marks the first time document contents alone have been used to produce a viable map of science. Millions of articles were semantically analyzed and their similarities-and the similarities of their respective journals-were used to create a two-dimensional visualization.

These types of maps can provide insights into the nature of science itself and are often used for high-level overviews of what is going on in science and how different disciplines relate to one another. The researchers' primary focus in this case, however, is visualizing individual user search results to support the search experience.

"This work is the basis for a service we are working on, where users submit a search and their results are overlaid on the map, providing context and the ability to identify and explore nearby relevant articles," says Glen Newton, NRC-CISTI researcher and the project's principal investigator. "Our next step is to work on implementing this process to improve and facilitate knowledge discovery for users."

This research is an important example of how Carleton and NRC-CISTI are working to advance research by providing access to high-value information services.  Developing these types of visualization and analysis tools will help users interpret and use STM information more efficiently to speed up the innovation process.

"Being able to build and navigate high quality, interactive maps of science from full text articles is a key outcome of leveraging Canada's emerging cyberinfrastructure," adds Michel Dumontier, project collaborator and associate professor in Carleton's Department of Biology and School of Computer Science. "Future integration with our Semantic Web enabled knowledge discovery platform will create new and exciting opportunities for Canadians to explore and discover interesting knowledge emerging from their investments in basic and applied research."

The paper for this work ("Semantic Journal Mapping for Search Visualization in a Large Scale Article Digital Library", Newton, Callahan, Dumontier) was presented at the Second Very Large Digital Library Workshop, at the 2009 European Conference on Digital Libraries (http://www.ionio.gr/conferences/ecdl2009/index.php).

Blog entry: http://zzzoot.blogspot.com/2009/07/project-torngat-building-large-scale.html