The Department of Homeland Security has a keen interest in the development of data-visualization tools that make it easier for intelligence analysts to extract key pieces of information from large volumes of structured and unstructured data. With terrorists taking advantage of the wide-open borders of the Internet and E-mail messaging groups to find recruits, raise money, and encourage jihad against the United States, the department needs a way to ferret out immediate threats as well as uncover sleeper cells and other dangers. So last year, it created the National Visualization and Analytics Center at the Pacific Northwest National Laboratory in Richland, Wash., with $2.5 million in funding to develop technology that combines data mining and analysis with visual-presentation capabilities to detect, assess, and reduce the threat of attacks, as well as respond to them.

The road to graphically rich, texturally detailed graphics that can help the center spot threatening trends--and might someday help businesses identify undervalued stocks or locate supply-chain bottlenecks--is being charted only now. Scientists and engineers have used sophisticated data-visualization tools on workstations for decades to design cars, model molecular compounds, and simulate the effects of tornados and other weather phenomena. But using computer graphics to model something tangible like airflow over an airplane wing is one thing. Adapting that technology to create interactive visual representations of less tan- gible data, such as terrorist activity or stock-trade patterns, has taken longer.

That's left much of the business world in the dust, even as data from Web sites, point-of-sale systems, and IT security tools rapidly accumulates in companies' data warehouses and the speed of business ramps up. People are left to glean trends and spot patterns from eye-straining text reports, floor-to-ceiling spreadsheets, and Excel's basic 2-D and 3-D bar charts, scatter plots, and histograms.

	(click image for larger view) Among the apps for Spotfire's DecisionSite data-visualization tools is oil-well production analysis.

"We're inundated with information," sums up the visualization center's director, Jim Thomas. "What we're trying to do is solve the problem of working with large volumes of often complex data." The center's work, which revolves around using spatial and graphical elements to convey more abstract data, has applications for businesses, too. The goal is to provide new tools for dealing with the huge digital libraries of the future, whether to detect terrorist threats or do genomic sequencing in the development of potentially lifesaving drugs. For its first partnership, the center is working with Stanford University researchers to develop methods to analyze network traffic patterns to graphically identify compromised systems used by consumers, businesses, and the government.

Confronting data overload, a growing number of companies are tapping into a new generation of data-visualization tools for businesses that they expect will let them make discoveries--find quality-control anomalies in a production run or the most-promising results in clinical drug trials--in minutes rather than days or weeks. Think 3-D representations of a day's worth of stock trades that appear as a mountain range with multicolored peaks and valleys signifying high and low stock prices, trading volumes, and other metrics. Some of the new tools superimpose data over geographic maps--a utility company might show electricity flow or power-substation status in the area it serves--or graphically represent a computer network or supply-chain system. And a comparatively new form of graphics called "treemaps" uses squares of varying size and color to visually represent data variables (see story, "Treemaps Rule: Even The Marines Think They're Cool"). Most of the tools that create these charts let users drill down into the details behind each one.

Pharmaceutical maker AstraZeneca plc uses data-visualization tools to better manage drug-development projects. Earlier this year, some 400 managers and employees at AstraZeneca began using Advizor Solutions Inc.'s data-visualization software to track project schedules, costs, and allocated resources. Pulling together data scattered across the company in spreadsheets and databases, the software provides color-coded bar and Gantt charts that help managers get a high-level view of the product-development process that project-management apps can't provide, project information manager David Scanlon says. "You're four clicks away from a dozen graphs" that provide snapshots of project activities, schedules, devoted resources, and costs, he says of Advizor's simplicity.

Yet the state of the art among most tools for business-data visualization doesn't hugely impress Scanlon, who has a background in genetics and molecular biology. Compared with the tools used for scientific work, visual-analysis tools on the business side are "clumsy and pretty rudimentary," providing simple dashboards rather than real interactivity.

(click image for larger view) American Water uses data-visualization graphics like this to aggregate data from security systems.

Chalk up the less-evolved state of the market to a few factors, including a lack of communication between the developers of data-analysis software and data-mining software, which are tailored for sophisticated users such as statistical analysts, and the vendors of early data-visualization tools. Startups, such as FYI, OmniViz, Spotfire, and Tableau Software, are bridging that gap.

OmniViz is a spin-off from the work done at the Pacific Northwest National Lab. Tableau got its start at Stanford University's department of computer science when the company's co-founder and chief technology officer, Pat Hanrahan, a Stanford professor--and now the researcher leading the university's work with the National Visualization and Analytics Center--was approached by the Department of Defense to develop a way to help it get meaningful data from a variety of databases.

Spotfire, which received seed money from In-Q-Tel, the CIA's technology incubator, counts among customers of its DecisionSite visual-analysis tool ChevronTexaco Corp. The oil and gas company uses DecisionSite to produce charts from geological testing data, which can take up as much as 200 columns and 30,000 rows on a spreadsheet, and well-production costs that help identify sites with greater potential for petroleum reserves. Now Spotfire has developed a tool built on DecisionSite that uncovers patterns in E-mail messages and displays the results in color-coded grids. The National Security Agency is using it to help identify terrorist messages. The software also has potential text-mining applications for businesses.