Refining the FirstGov map

Visitors to the federal government's Web portal literally do not know what they're missing: a trove of information — compiled by the U.S. Geological Survey, the Census Bureau and other agencies — that is fully realized only when linked to a map.

FirstGov officials understand the value of geospatial data but haven't yet found a way to make it available through FirstGov.gov, which was engineered to search and display text documents. That may soon change. The FirstGov team, working with the Federal Geographic Data Committee, has asked the geographic information systems (GIS) industry to submit ideas for unlocking geospatial data stored in agency databases.

"A lot of information in government has a geospatial component to it. People like to see information brought up graphically," said Keith Thurston, assistant to the deputy associate administrator for the Office of Governmentwide Policy at the General Services Administration, which manages FirstGov. "What we're looking for is industry to make proposals on the best way to approach this."

FirstGov's search engine combs through the text of some 27 million Web pages with URLs ending in ".mil" or ".gov," looking for keywords entered by users. However, it cannot return search results relevant to a user's location — what is known as georeferencing — or produce maps illustrating geospatial data.

FirstGov officials initially hope to add a federal office locater to the portal, Thurston said. For example, if you live in Manassas, Va., you would be able to type in that city and have a map show you Social Security Administration and Internal Revenue Service offices nearby, similar to the driving directions users can obtain through MapQuest.com Inc.

FirstGov's search engine is powered by a fairly standard spidering program, or Webcrawler, that searches for government Web sites and copies and saves them in a database every two weeks. The search engine then looks for search terms entered by users within the text of all those saved Web pages.

"The biggest difference with any type of GIS system is that the engine itself can't determine where the GIS aspects of a target Web site are," said Dave Binetti, president of the Fed-Search Foundation, the nonprofit organization that manages FirstGov's search technology. "You have to have some sort of tagging architecture so that the sites that have geographic components can indicate that to the engine."

The Open GIS Consortium Inc., an industry standards group, is developing such a technology, called the Geography Markup Language. GML is based on the Extensible Markup Language, an increasingly popular technology used to encode information for Web pages or online transactions.

Members of the GIS community are eager to see FirstGov develop its GIS potential. By some estimates, some 80 percent or 90 percent of all data stored by the government has a geographic component to it.

"The breadth of content available from the government is huge," said Dean Kensok, a spokesman for GIS software vendor Environmental Systems Research Institute Inc. "It's really almost more an administrative challenge to get all that content integrated than a technical challenge."

In addition to georeferencing data on Web sites, Kensok said FirstGov could make actual GIS data more available to FirstGov users. For example, a community mapping application might cull data on toxic and flood hazards and census demographics when users type in their ZIP codes.

The Federal Emergency Management Agency, USGS, Census and other agencies provide mapping services individually, but there isn't a service that can map all that data simultaneously, Kensok said.

GIS systems are specialized to fit the needs of individual organizations and federal agencies, GSA's Thurston said. FirstGov's challenge is to make proprietary data accessible to ordinary citizens using PCs and Web browsers.