Data miners

Agencies in the homeland defense era are turning to business intelligence tools to mine and tag nebulous data, drill down to details and spot unsuspected relationships.'We are just beginning to make full use of data mining strategies,' said Gregory D. Kutz, director of financial management and assurance at the General Accounting Office, and a longtime investigator of fraudulent purchase card use. Kutz recently told a congressional panel that GAO believes data mining will become one of its most important tools.As more business intelligence products appear'especially at agencies engaged in the business of intelligence'so do doubts about their invasiveness. Opponents of data mining believe that giving agencies a peek into citizens' daily lives will violate privacy and civil rights. Three senators cosponsored legislation to place a moratorium on data mining by the Defense and Homeland Security departments, until Congress has fully reviewed the controversial Terrorism Information Awareness project.Critics say TIA, which has a $137 million budget for fiscal 2003, would let intelligence agents rifle through traffic tickets, credit card statements, medical records and other personal information without enough oversight.The House Government Reform Subcommittee on Technology, Information Policy, Intergovernmental Relations and the Census last month turned a spotlight on the Defense Advanced Research Projects Agency's TIA, the Transportation Security Administration's Computer Assisted Passenger Prescreening System II and the FBI's Trilogy network.Data mining 'refers to the ability to work with larger amounts of data, at faster speeds, in ways that were previously not possible computationally due to size or speed limitations,' said Steve McCraw, assistant director of the FBI's Office of Intelligence, in testimony before the subcommittee. The intelligence office will use the technology to match fingerprints against large databases to pinpoint high-risk employees.'In recent debates,' McCraw said, 'people have begun to use the term data mining as a shorthand reference to the specter of abusive searches through data on innocent private citizens''two different concepts, he said.'Clearly, the events of Sept. 11 heightened awareness and caused us to examine ways to share information,' said Bob Dix, the subcommittee's staff director. 'Data mining is one way of doing that.'Other agencies use business intelligence to make sense of projects that get far less public deliberation. GAO has used the IDEA data analysis tool from Audimation Services Inc. of Houston to shorten some of its purchase-card fraud investigations by weeks, officials said. IDEA can scan millions of bank records to sort improper charges by amount, seller, buyer, date or location.'You couldn't possibly print out everything and do it visually,' said Alana Stanfield, GAO assistant director of financial management and assurance.Another business intelligence tool has done the work of three full-time employees in the Environmental Protection Agency's Superfund program. They used to spend a month drafting a dozen waste site reports. Now a server nightly updates 55,000 waste-site Web pages via the eFacts portal from KeyLogic Systems Inc. of Morgantown, W.Va.'We're putting information in people's hands that didn't have it before,' said Erin Conley, a Superfund program analyst at EPA. 'They can do more and make better decisions.'The $600,000 eFacts portal lets EPA intranet users drill down from a site's general information to the amount of money being spent on it and the timeline for reconstruction projects. 'Now all the regions are on the same page,' Conley said. 'Everyone can look at the same data.'An office at the Bureau of Labor Statistics plans to start using new data metrics software this fall to assess how field economists collect employee wage, earnings and benefits data from thousands of companies. The Office of Compensation and Working Conditions will use Cognos Metrics Manager, from Cognos Inc. of Ottawa, to send top managers scorecards showing how well regional and national teams of economists are meeting milestones and gathering quality data.'It will help us spot troubles before they develop,' said Randy McLin, computer specialist for the office's IT arm. Until now, he said, 'Different people had different measures for where we were.'At the FBI, automated data analysis tools will power a new counterterrorism database, letting 300 bureau analysts dig through more than a billion documents and share information with other intelligence agencies.The tools, ClearTags and ClearResearch from ClearForest Corp. of New York, will search for patterns in terrorism-related intelligence collected from multiple sources into a centralized data mart on the bureau's modernized Trilogy network.