With a few keystrokes, a researcher can call up the locations of the streets of any Connecticut town. She can learn the types of soils those streets are built on and the ages of the people living in the houses nearby. She can find wetlands and ground cover data to see where threatened species live, and where pines give way to asphalt. She can create maps with this data, and layer the maps together to see how homes and forests intertwine. She can analyze them to find the best site to build a new factory.
It seems like magic, and it is: MAGIC, the Map and Geographic Information Center at the Homer Babbidge Library of the University of Connecticut (UConn).
The map library at UConn comprises an extensive collection of over 150,000 paper maps, to which thousands are added each year. It serves as a federal depository, holding all maps published by the federal government, including the United State Geological Survey (USGS) topographic maps. Its collection holds historic maps, demographic maps, transportation maps, and New England town plans dating from circa 1850.
But over the past five years, an increasingly important component of MAGIC has become its computer database, which supplies geographic information about Connecticut, in categories that range from ethnic distribution to hiking trails to town boundaries to drainage basins.
Initiated by map librarian Patrick McGlamery in 1992, these digital files were set up to allow users to take advantage of GIS (geographic information system) analysis, a broad and potent form of investigation that, over the past decade, has grown into a multi-billion dollar industry with hundreds of thousands of practitioners worldwide.
Although it is map-based, a GIS is only a distant cousin to the common paper map. It is a combination of software and database with which users can perform a wide range of tasks. With a GIS, users can combine information about highways with population density and income data to find the most profitable site for a new shopping mall. They can layer maps of childhood leukemia clusters with maps of toxic waste sites to seek out possible correlations. With a GIS, users can determine bus routes, and analyze the equity of political voting districts. They can learn where endangered species might be located, find the most suitable sites for schools or factories, develop emergency evacuation routes, plan for future land use, and track changes in the growth of forests.
Unlike most databases, in a GIS each piece of information possesses a geo-reference: it can be pinpointed with respect to the earth's surface according to its longitude and latitude. Data might include single 'points,' like a house, a well, or a car accident; 'lines,' such as roads or rivers; or 'polygons,' continuous, boundary-enclosed areas like ponds, census tracts, or parking lots. Non-spatial attributes, such as the age or depth of a well, the type of injuries in a car accident, or the population density of a census tract, are also included.
Because every piece of GIS data shares this same longitude/latitude grid, it is possible to juxtapose these widely diverse categories of information, performing analyses and finding correlations that would otherwise be obscure and unwieldy.
In Connecticut, GIS technology has been used for a wide variety of purposes.
Even high school students are learning to use this method of analysis. In a summer program at UConn, 52 middle and high school students from eleven Connecticut towns used GIS for land planning. Working with data about the town of Mansfield, the students analyzed town boundaries, hydrography, transportation, soils, drainage basins, public lands, rare and endangered species sites, and other data, using modeling techniques to allocate town land effectively for residential, commercial, industrial, and recreational use.
Developed in the 1960s by the Canadian government as a way to provide a land use inventory for their country, GIS has only within the past decade begun to come into its own. In large part, this has resulted from the advent of personal computers powerful enough to manage the vast amounts of data required by a GIS, and from the development of spatial databases by various arms of the US government. Completed in July, 1995, this network of 24 satellites uses triangulation to determine the longitude, latitude, and altitude of a GPS receiver on the earth's surface. For GIS data, it provides greater precision, allowing a user, for example, to confirm the exact location of images received through LANDSAT satellite or aerial photography.
Much GIS data is available commercially. But the information on MAGIC is something of a different animal. MAGIC is organized around the belief that information collected by federal and state agencies should be available to everyone. MAGIC tries to make all appropriate government-compiled data accessible to the public as easily and as widely as possible. "MAGIC," says map librarian Patrick McGlamery, "is driven by the notion of public domain."
McGlamery conceived the idea for MAGIC in the 1980s, when the Bureau of the Census first began to release digitally formatted census data, known as TIGER files, through CD-Roms. The files contained not only all the demographic information collected by the Bureau, but, for the first time, all the streets in the United States, in rural as well as urban areas, with their address ranges. [That is, street numbers were attached to the roads, so that researchers could know the locations of specific addresses.] McGlamery, with an LSCA (Library Services and Construction Act) grant, purchased the CD-Roms for the map library. But the census files, as they were, were not easily usable by researchers.
"When TIGER first came out, it was just streams of ascii-formatted data, without even carriage returns at the end of each record." McGlamery said. The files could not be used until they were converted to a more accessible format.
While most libraries chose to purchase already-converted files from a private company, McGlamery decided that, for MAGIC, library staff would do the work. "I felt it was important to keep the data in the public domain. If libraries process the data themselves, they're able to give it away, the way we are. But if [a private company] touches it, then they own the copyright." In such cases, the libraries are allowed only to distribute maps made from the data, rather than the actual data itself. A universally accessible database like MAGIC would not be possible.
By 1992, McGlamery began to incorporate data from other sources, beginning with the Department of Environmental Protection (DEP). His concern remained that of releasing government information to the public. "While there has been some legislative language that indicated that 15 copies of anything published by a state agency would go to the state library, when information began to move to an electronic format, there was always the argument that the information had not been actually published."
The library and the DEP worked together to develop a protocol for releasing electronic data, and now MAGIC serves as a de facto depository for appropriate DEP material.
In addition to providing information from the Census Bureau and the DEP, the constantly growing database gathers files from the state Department of Transportation (DOT), the US Geological Survey (USGS), National Fish and Wildlife Service, and the National Soils Conservation Service. Some Connecticut towns are also beginning to supply local data.
MAGIC provides information on housing, economic, and labor statistics; on historical populations from 1774 to 1990; and on age ranges, education, and quality of life within the state. It offers data on railroads, streets, town boundaries, drainage basins, hydrography-including streams, ponds, marshes, tidal flats, ditches, and channels-and 23 categories of land use/land cover, ranging from roof to pavement to high density or low density residential to deciduous or coniferous forest to pasture to coastal marsh to forested or non-forested wetlands. The land use data is derived from LANDSAT Thematic Mapper Satellite Imagery information.
The information resides on MAGIC as digital data, not as maps. However, with GIS software, it can be viewed in map format. The data are offered, in most cases, at quadrangle, town, county, or state levels in scales that range from 1:100,000 to 1:24,000. The data arrive at MAGIC in a variety of formats, including ascii, digital linegraph, digital elevation, and so on. In order to make them accessible, the library staff transforms them into more universal formats, adding spatiality (latitude and longitude) where necessary. All MAGIC data use the Connecticut state plane coordinate system: that is, despite the earth's roundness, the entire state is treated as if it was flat. This allows it to match the flatness of a computer screen.
Users must have GIS software in order to read MAGIC files, which are made compatible with three common GIS programs: MapInfo, a desktop mapping program commonly used by businesses; AutoCAD, often used by engineers; and ARCInfo, which is widely used by researchers. For users who lack access to these programs, MAGIC provides a link to free, downloadable, compatible GIS software.
The files on MAGIC are widely used. In the first eleven months of 1996, 12,986 files (over 2 billion bytes of compressed data) were downloaded, about two-thirds of these by individuals not connected with educational domains.
MAGIC will serve as one leg of the newly established University of Connecticut Center for Geographic Information and Analysis. The center will, according to McGlamery, "put a mechanism in place whereby the university can begin to provide some of the services it should provide." Headed by UConn geography professor Robert Cromley, the center will serve three functions.
GIS is so widely used, says Professor Cromley, that there are already other centers on campus that rely heavily on GIS activities. But he feels that because of the rapid growth in the field, and because of the technical complexity of managing geodata, the expertise provided by a Center is clearly needed. All other New England state universities, he points out, already have some form of a GIS center.
Professor Cromley hopes that the center will serve as an umbrella for GIS-related activities. "It has a mission," he says, "to foster studies involving geographically referenced information in the state of the Connecticut."
In addition to adding more data, McGlamery is making MAGIC easier and more effective to use. With a $40,000 FGDC (Federal Geographic Data Commission) grant, he is completing an online written description of each file. These metadata-data about data-include information such as who collected the data, when, the methods used, the coordinate system, and the scale. It will adhere to the spatial metadata content standard developed by the FGDC in 1993. Metadata provide users with the information needed to evaluate the data, allowing users understand its limitations: for example, that data collected at a scale of 1:100,000 should not be used to plan a minute structure like a culvert.
The metadata descriptions will also make it easier for users to locate information. Currently, because MAGIC files consist primarily of numeric data, users can find information only through the Web site's graphical interface. To find a piece of data, the user must first know which file it is in. For example, if a researcher wishes to look at a particular farm or road, he must know the town or county in which it is located. Metadata descriptions will circumvent this requirement by permitting keyword as well as spatial searching. Keyword searches should be available by the end of 1999.
This metadata project is being conducted in conjunction with an NSDI (National Spatial Data Infrastructure) program to develop a means for locally-collected geodata, like that available on MAGIC, to be shared throughout the country. The NSDI is in the process of organizing a nationwide federation of geodata servers that will collect, or at least catalog, metadata about their area, allowing it to be searched by a search engine. MAGIC is in the process of becoming a node-a server-in this program, which will greatly increase the range of information available to MAGIC users.
"If you're interested in soils data for Connecticut," McGlamery explains, '[through the NSDI project] you could query the distributed node, and it would find data here on this server in Connecticut, of course. But it might also find it in some other place: in Wisconsin, maybe, with a researcher who's doing glacial studies that relate to Connecticut soil."
In fact, McGlamery foresees a time when it may no longer be necessary for MAGIC to contain data at all. Instead, information would reside with the agencies that collect it, and MAGIC's function would be to maintain a metadata catalog providing links to the agencies, and to other NSDI nodes. Through those links, the MAGIC Web site will be able to provide its users with geodata not only about Connecticut, but about the entire hemisphere.--Karen Miller, freelance science writer.
To contact map librarian Patrick McGlamery, call (860) 486-4589.
Back to TOC page