Search Show Offs


      Bookmark and Share

It seems that my clients have been conspiring against me again. Yes, I already know that they coordinate all their rush requests for two hours before I have to leave on a trip. But now they are sending me more amorphous questions than usual. Instead of being asked to look for information on trends in the consumer DSL market, for example, I get projects on the growth of the political polling market or measuring the return on investment of executive training and development programs. I find that my searches, particularly in the value-added online services, are a lot broader, which means that rather than wading through 15 or 20 records, I have to slog through 200 or 300 to find the answers I need.

Keep in mind that I use whatever tools I can when reviewing search results, including Factiva.com's relevance ranking and LexisNexis' FOCUS feature. But I am still faced with making a purchase decision based on little more than a bibliographic citation and the fact that one article appeared three screens ahead of another, similar-sounding article. It just feels so…well, 20th century.

This has triggered an interest in visual informational displays. As anyone who has read Edward Tufte's books—particularly The Visual Display of Quantitative Information—knows, presenting information in a meaningful, efficient, and elegant way is an art. What I find particularly compelling are Tufte's thoughts on how to represent three dimensions of information on a flat surface or, in the case of electronic information, on a screen. (For examples of occasionally good, efficient displays of data, check out the "Infoporn" feature in each issue of Wired magazine.)

What online services and most search engines currently offer in displaying search results is both visually dense and information-poor. So I have started keeping my eye on trends in data visualization tools. These tools work best with homogeneous data sets; they're most powerful with information that is built in fields or that has some sort of controlled vocabulary, rather than Web search engine results, for example. But for structured data, such as you would find in most econtent sites, data visualization can rock.

Check out Plumb Design's Visual Thesaurus, for example (www.visualthesaurus.com). Type a word into the text box and you'll retrieve an almost intoxicating display with the word and all its various meanings and related terms. Speaking of intoxicating, try the word loaded. You'll see little branches spring forth for the concepts of affluence, laden, drunkenness, charged with ammunition, and so on.

Other companies working on the challenge of visual display of information include Inxight (see some demos of their VizServer product at www.inxight.com/products/vizserver/); Antarctica Systems (which even describes a library online catalog application at its site, www.antarcti.ca); and TheBrain Technology (see www.webbrain. com for a cool demo of mapping Web search results). Chris Sherman, co-author of The Invisible Web, introduced me to Map of the Market, at the SmartMoney.com site. This ingenious page displays the current status of the stock market and the relative market cap of individual companies and industries, at a glance. You can see what's up or down, and mousing over a specific square gives you details on that company's current stock price and activity.

These data visualization tools work best for small bits of information—finding relationships among words, drilling down through organization charts, or rooting around in library catalogs. But the WebBrain application, in which entries in the Open Directory Project (www.dmoz.org) are displayed in a visual format, got me thinking. Why not use one of these tools to enable searchers to do their own "more like this" browsing of search results in an econtent site? Say I'm looking for articles on the ROI of executive training. I've got 300 possibly-relevant articles. I find one that looks promising, so I click on the keywords associated with that article. I get a display of those words, the articles from my search associated with each of those keywords, and links to related keywords that didn't appear in that one relevant article I looked at. Now I expand the vision of one of those related words, see the articles in relevance-ranked order that look good, and repeat the browsing process. I'm not just doing a linear browse of article titles like we info pros have been doing for decades; I'm exploring the universe of content in several directions (or dimensions) at once.

This technique is a real conceptual leap for many researchers. We're used to Boolean logic, step-by-step processing of searches, and chronological display of search results. To think of displaying information in a dynamic, multi-dimensional format is a little like considering the transition we non-Mac folks made from MS-DOS to Windows. It's a completely new, and much more powerful, way of looking at information. The question now is whether the econtent companies can develop the tools to let us info pros see that content in three dimensions.