Biologists and chemists at Infinity Pharmaceuticals deal with a daily delicate balance: developing therapies to treat cancer and related conditions, and sharing analysis in a timely manner. Scientists lacked a common analysis environment, exchanging data for cancer drug development at a crawl. Biologists and chemists tracked most of their work in Microsoft Excel, first copying and pasting results into text files, then manually punching in calculations to view data and graphics in Excel.

“When a lot more work is done manually, you have a tendency for human errors,” says Jeff Reed, Infinity’s software engineer analyst. Scientists created their own spreadsheets, resulting in multiple versions floating around, slowing discovery cycles and bogging down production. Infinity selected Spotfire Decision-Site software to help it make more accurate decisions and share knowledge among scientists.

“A ton of data goes back to the scientists, and it’s a quick and easy way to visualize all sorts of data,” Reed says. “You can take 50, 60, 70,000 rows of data, bring it up and show scientists what experiments look like.”

DecisionSite works with Infinity’s Oracle 9i database, pulling information and sending it back to scientists in a visualized format—bars, graphs and pie charts. Infinity created a Web application to set up experiments, and another application loads into Oracle 9i to conduct calculations.

The discovery cycle took a couple of days, but scientists now run an experiment and receive results in less than an hour, Reed says. In addition, DecisionSite shaved up to 4 weeks per year off the project team data processing and decision cycles.

DecisionSite is a “visualization warehouse, and you always have the same amount of data everyone is looking at…there are no apples- and-oranges comparisons,” Reed says. Scientists run, load and view data without making requests to Infinity’s IT division. DecisionSite typically identifies 90 to 95 percent of drug development issues, according to Reed, allowing scientists to spot problems quickly and move onto the next experiment.

Visualizing business value
Data visualization software, a niche within business intelligence, allows companies to drill deeper into content, viewing and interacting with information in ways Excel and static charts can’t match. Enterprises, including healthcare, aerospace, consumer electronics, pharmaceutical and computer hardware, jumped on data visualization quickly to identify trends and problems in inventory, sales, customer behavior and other areas, improving their grasp on operations.

For example, Briggs & Stratton implemented SAS Institute Enterprise BI Server to identify trends and anomalies across its product lines, answering such questions as: What is the inventory rate compared to sales? Why are sales up? What is the fastest-growing product?

Briggs tracked potential engine problems, preventing months of faulty production, saving $4 million per year in warranty payouts. (See related story, “Briggs & Stratton mows down data with visualization.”)

Some businesses get along fine using Excel and PowerPoint to depict their data, while others keel over at price tags that average hundreds of thousands of dollars for data visualization systems. Data visualization is catching on, albeit slowly, especially with companies hesitant to spend IT budgets on a niche product. Guy Creese, an analyst at Ballardvale Research, says companies don’t see the possibilities of data visualization. “Its past history was…it is eye candy; it’s taking data readily available and dressing it up pretty,” Creese says.

He acknowledges businesses rely on reports, but data visualization helps them see patterns, especially subtle ones often overlooked. “They’re good for different things; reports are more detailed oriented,” Creese says, “Visualization allows you to break down these barriers and look at [data] in a different way.”

Data visualization fills a void in BI: All BI vendor products include basic charting, data discovery and information delivery, but data visualization offers these features, as well as visual query, multiple linked images, dynamic data content and presentations of large, complex multidimensional data sets, according to Forrester Research. (See related story, “One way to actively look at data visualization.”)

Advances in Web-based graphics and applications are making data visualization software more affordable, costing tens of thousands of dollars, compared to a system’s six-figure cost, Forrester says. Revenue for the visualization market is hard to come by—the overall BI market racked up $5.5 billion in 2004—with a mish-mash of players vying to enter the action.

BI vendors including Cognos and Business Objects offer data visualization capabilities in their products, while Antarctica Systems Visual Net and DecisionSite are development environments that create analytic applications with visual, interaction and collaboration features. Corda Technologies PopChart, Visual Mining NetCharts and others provide static to moderately interactive chart-centric products. Data mining tools, such as SAS Institute Enterprise BI Server, generate decision-tree data structures, while testing mining tools such as Entrieva SemioMap and Inxight VizServer find relationships among topics in unstructured text documents.

Data visualization offers companies, especially decision makers, a view to complex issues, according to Keith Gile, a Forrester analyst. “It simplifies and focuses attention on…what’s important,” Gile says. “You need millions of rows just to come up with a tendency, and it’s hard to see [that] in a spreadsheet.”

IT views the wealth
Businesses aren’t the only users getting the benefits from data visualization software. IT departments are spared from writing multiple queries and creating numerous reports. Many IT departments write a query to create reports for users, and if those users want more details, IT writes another query, taking time from important projects. The process isn’t made any easier when IT tries to pull this information from disparate sources.

“Data visualization on the Web is so much easier,” says Wayne Salter, an integration services senior analyst at Chick-fil-A, which implemented Corda PopChart to assist in tracking operators’ drive-through sales and a restaurant-level scorecard for about 1,200 restaurants in 38 states.

For example, Chick-fil-A uses custom Java code to extract data from its J2EE-built enterprise application and pulls it from their data warehouse for drive-through performance reports. “It really brings those projects we passed on because we didn’t have enough staff; it brings some of these projects into the realm of possibility. They are now within reach.”

IT staffs also use data visualization software to help with their own projects. DePaul University implemented Antarctica’s Visual Net software for project management. DePaul’s IT staff had basic, tabular reports showing a project’s status, resources used and other information, but something was missing. “Visual metaphors map well into project management problems,” says Vince Kellen, DePaul’s vice president of information services.

Kellen selected Visual Net so DePaul’s IT staff could access it online from anywhere. Data is stored in the project management database, which runs on Microsoft SQL Server. Visual Net, which runs on a dedicated server, reads information from other servers and then delivers it to IT.

Visual Net allows Kellen and his IT staff to drill down and find information related to project management, such as estimates about the amount of staffers required for a project. Kellen anticipates using Visual Net for financial, budget and expense tracking.

“It actively manages projects, and we tend to see these patterns, no question,” he says. If functional requirements get bogged down, DePaul’s IT staff can take action and make any necessary adjustments.

Sidebar: One way to actively look at data visualization
Case Study: Briggs & Stratton mows down data with visualization
Case Study: Sun banks on data visualization to win customers