A Picture Is Worth a Thousand Charts

New software visualizes data in ways that give a much richer image than typical tools

If you want to know how the markets did today, you could try to make sense of the tickers on CNBC (GE ) or the charts and tables on any number of Web sites. Or you could look at the Map of the Market on www.smartmoney.com. There you will see blocks of colored rectangles grouped by industry. Each block is a listed stock. Its size indicates the company's market capitalization, and the color indicates price change. Clicking on a company offers more detailed information about the company and its securities.

The map, which conveys lots of data at a glance, is an example of a technique known as information visualization that is making its way into the mainstream after a long gestation in computer science labs. The idea is to replace the rows and columns of numbers or the graphs and charts that have long been used to present data with a much richer visual picture. Visual presentation isn't a substitute for numbers and graphs, but an adjunct to them. "Information visualization answers questions you didn't know you had," says University of Maryland computer scientist Ben Shneiderman, who developed the underlying technology that smartmoney.com uses. (I urge you to check out the Web sites for demonstrations of the products I describe here--their visual nature proves a picture is worth at least 1,000 words.)

For the most part, visualization tools are cropping up as part of the analytical software packages that are used to mine data at large companies. However, there are stand-alone products that use the technique to help users make visual sense of what otherwise might be an overwhelming sea of data. One approach is called Grokker, downloadable as a $99 "preview edition" from www.groxis.com. Grokker is a Web search tool that instead of responding to a query with pages and pages of links, shows the result as a series of circles. A big circle represents all the results. Circles within it represent subtopics, with the size of the circle indicating the amount of data it contains and the color coding the relevance of the data. Each inner circle expands as you click on it, revealing additional detail, until you click on the innermost circle, which then opens a Web page.

The Grokker approach works best on vague queries where you don't really know enough to construct a precise search request. For example, a Google search on "data visualization" produced 110,000 results. Grokker produced 13 major categories, including one, which was what I really wanted, that gathered the results dealing with decision support software. As underlying search tools, Grokker uses Northern Light (www.northernlight.com), which offers a lot of results you must pay to see, or the Open Directory Project, a free, cooperative effort organized by Netscape (AOL ) to build a massive catalog of the Web.

Most of the visualization tools are designed to extract information from existing databases, but MindManager from Mindjet (www.mindjet.com) lets you jot your musings about a subject and assemble them into a "map" of ideas. The result is something like an outline, with topics and as many layers of subtopics as you want. But instead of an outline's linear organization, MindManager ($199 for the business edition, with a 21-day free trial available) lets you put the information anywhere you want as you build a tree of ideas. Simple tools let you collapse branches to see the big picture or focus on an expanded branch to drill down into the details. Items can be color-coded or tagged with icons for identification. It is also easy to connect the data in MindManager into more conventional, structured tools. For example, to-do list items emerging from a brainstorming session can be exported either as Microsoft Outlook tasks or as checkpoints in a Microsoft (MSFT ) project timeline.

For the past 20 years, the vast data-management power of PCs has been used mainly to emulate the paper forms and reports of the past. Data visualization literally changes how we look at information, and it will take some getting used to. But, as Shneiderman writes in Leonardo's Laptop, a newly published plea for a more humanistic approach to computing, "most people depend on visual input for much of their understanding of the world around them." In a world where our problem is frequently too much information, these techniques can be a big help.


    Before it's here, it's on the Bloomberg Terminal.