Factiva's very existence depends on providing its subscribers with the right business news and information drawn from 9,000 sources, including Dow Jones Newswire, the Financial Times, The New York Times, Reuters Newswires, The Wall Street Journal and 350 million archived documents in 700 different formats. To compete with LexisNexis and Dialog, Factiva must provide effective search. "When you're making a critical business decision, you want to get all and only the relevant content to that decision," says Alan Scott, chief marketing officer.
The company recently upgraded its Factiva Search component to version 2.0, using tools from Fast and Inxight Software and rules-based feed technology from Teragram. The company simplified the user interface and added visualization while preserving its taxonomy.
Factiva subscribers are now presented with a typical search box and receive snippets of results with the search term highlighted. That's nothing special by itself, but a "discovery pane" on the right side of the screen holds a set of unusual search-visualization tools that help users find related bits of information that they might not have known to look for. A bar-chart time line shows when the topic has been most and least in the news. Click on a bar and you get the specific results for that time period. Clusters identify relationships between the search term and other words and phrases that occur within a specified proximity, such as within 10 words of the initial search term.
Factiva converts its 200,000-per-day incoming documents to a uniform XML format using autotagging technology. Editors then check the tags on each story to make sure they fit the company's taxonomy. They have the best of both worlds: some automation to reduce manual labor and human intelligence to review and revise the results. Factiva's content and search technology are UIMA-compliant, which means other UIMA applications will be able to use it.