GIS Devotees Map Out Their Turf
GIS Devotees Map Out Their Turf<@VM>Mapping the Big Apple<@VM>Breast Cancer Study
By John MakulowichWhile not a torrid topic of everyday chat on the Internet, the subject of geographic information systems enjoys a
devoted following among diverse professionals, from analysts mapping crime to environmental agency executives managing the map preparation of major cities to medical researchers seeking
the causes of breast cancer.
GIS amounts to a computer tool for analyzing geographic-based events by integrating common database operations, such as query and statistical analysis, with maps. A GIS stores information in a database as a collection of different layers that can be linked together by geography.
One of the more popular uses of GIS is to map crime data, the subject of a recently published research report by Keith Harries, "Mapping Crime: Principle and Practice." Available on the Internet as a PDF file (www.ncjrs.org/html/nij/mapping/pdf.html) and funded by the National Institute of Justice, the research agency of the Justice Department, the study serves as an introduction to the science of crime mapping for police officers, crime analysts and others interested in visualizing crime data through maps. It includes more than 110 GIS maps, references and an appendix of World Wide Web resources.
As noted in the study, about 13 percent of law enforcement agencies are using GIS regularly to analyze their crime problems. This number is likely to increase significantly as more and more agencies begin using computerized mapping to identify and solve their crime problems.
Said Harries, a professor at the University of Maryland, Baltimore County: "One main trend in GIS is toward greater integration of technologies and databases, of bringing diverse data sets together on issues that cannot be solved from one perspective."
Some interesting uses of GIS techniques and tools can be seen in Web sites maintained by local police departments, such as the Berkeley Police Department in California (inberkeley3.ci.berkeley.ca.us/annual_rpts/annual.html) or the Cambridge Police Department in Massachusetts (www.ci.cambridge.ma.us/~CPD).
For example, the Berkeley site offers a Crime Map hyperlink, where one can view statistics in a specific census tract. Statistical data is available from 1990 to 1998. Thus, Census Tract 22 with a population of 3,169, or 3.1 percent of the city's total, accounted for 2.2 percent of major crimes in 1998. Included in this category were aggravated assault, arson, auto theft, burglary, homicide, rape, robbery and theft.
On the Cambridge Police Department Web site (www.ci.cambridge.ma.us/~CPD/crime/index.html), users can view data covering annual and quarterly crime reports and neighborhood and business district profiles, as well as monthly statistical updates, patterns, trends, problems and weekly crime reviews.
For the period from Jan. 14 to Feb. 13, you can view a map (www.ci.cambridge.ma.us/~CPD/crime/sr.html) showing a weekly review of street robberies, which covers bikejacking, carjacking, home invasion and purse snatching.
At the high end of the bandwidth spectrum, for those who can travel at that speed, you can view the virtual reality site maintained by the Metro Nashville Police Department in Tennessee (www.nashville.net/~police/VRMovies/ index.htm). The police use the technology to link views at violent crime scenes. Thus, jurors can see where the defendant, witness, officer or subject was and click to go to other points visible in the movie. The links bring up sounds, pictures or object movies, which can be displayed on a large-screen television for the court.
On another dimension ? the level of an entire city ? is the work managed by Alan Leidner for New York City. The director of citywide GIS for the Department of Information Technology and Telecommunications, Leidner oversees the recently completed project, begun in 1997, to produce the city's first photogrammetric land base.
Photogrammetric is an extensive collection of aerial photos that includes topographic contours and spot elevations. It serves as a common, uniform base map for other layers of information as well as for the integration of data sets.
According to Leidner, the city has taken delivery of the physical base map from the vendor. The digital map of the city has more than 50 features, including building footprints and spot elevations. The total size of the file comes to about 50 gigabytes.
The IT department used an Oracle 8i database with its spatial option to produce a data warehouse to accommodate both spatial and non-spatial data quantities. The citywide GIS data warehouse will be built on an IBM ES9000 OS390 mainframe. New York expects citywide intranet services to be operating within the next three to six months.
One lesson Leidner learned over the two-year project was that everything cannot be covered in the specifications when the project is first scoped out.
"The only real way to do GIS in a local government environment is to think comprehensively," he said. "You cannot do GIS for just one project. You have to consider all the data and how it becomes integrated in order to prepare an accurate land base. You have to realize that all the data with a geographical component can be integrated into a GIS utility."
At the time the project started, Leidner was GIS director for the city's Department of Environmental Protection. That department is responsible for protecting and maintaining the water supply for 9 million city and upstate New York residents. It also oversees the operation of three watersheds that cover 2,000 square miles, 19 reservoirs and three controlled lakes.
Further, it treats waste water at 23 water pollution control facilities in the city and upstate watershed, and maintains water and sewer systems made up of nearly 6,000 miles of underground piping and 1 million building connections. Finally, the department enforces city air, noise and pollution control codes and is responsible for meeting air quality standards, enforcing asbestos abatement rules and responding to hazardous material emergencies and toxic site remediation. To manage these assets, the department has been using GIS since the 1980s.
Leidner said the recently completed project really took off when it got a champion in Mayor Rudolph Giuliani's office of operations, which brought it to the attention of a citywide audience.
"The sooner and more forcefully you can get support from different agencies, the easier will be your job. Otherwise, it will be an uphill battle all the way," Leidner said.
A critical key to success is getting a good GIS consultant, one who knows his or her business and can offer soups-to-nuts, cradle-to-grave consulting to cover the smallest detail, such as the specifications for aerial photography.
Part of the process of producing the photogrammetric land base involved two sessions of aerial photography at different scales, the first taken in spring 1996 and the second in March 1999. For mapping accuracy, airborne and Global Positioning System control surveys of photo-identifiable points were used, Leidner said.
What made the project somewhat easier was that the information technology involved was starting to mature: the software (such as Oracle's Spatial database), the hardware (such as faster servers) and the telecommunications interface (such as the Internet and the Web).
Plans call for New York's Hunter College, which enjoys a world-class department of geography, to do base map maintenance. Another procurement is in process for implementation over a four-year period to build a repository with new data layers.
Leidner, who talked with colleagues in forward-looking GIS communities such as New York's Nassau County, Philadelphia, San Diego and Seattle, said once you build the infrastructure, the costs to maintain the system are small comparatively. In those cities, central repositories of GIS data, often made up of more than 100 layers, have been created and made available to agencies as well as private firms and the public. New York City now is exploring the architecture required to build such a repository.
"While you can spend plenty of money building GIS in individual departments, when you unite the effort, not only is it less expensive, but the coherent program will be more valuable to the city," said Leidner.
In his estimate, the commitment required for a large jurisdiction to build central maps that "will get you a long way down the road" is about $5 million. The real benefit of such projects, he said, is to make public and private organizations operate more efficiently and increase profitability. This, in turn, increases the tax base for local government.
For the future, there are plans to develop applications that offer access to the land base and other data through the Internet and desktops. The city also is studying the potential of the Internet to make GIS information available.
Specifically, the city is interested in GIS applications that allow the public to do business electronically with the Web without needing to visit a city office or fill out paper forms for manual processing.
Away from the tall buildings, shadows and canyons of New York City, you can still find GIS at work in the microscopic world. One example is the project pursued by AverStar Inc., Burlington, Mass., under contract to the National Cancer Institute (NCI).
The company is producing a prototype GIS for breast cancer studies as part of the Long Island Breast Cancer Study Project. The first phase calls for development and delivery of the GIS system, while phase two is for system maintenance and data expansion to respond to research needs. The contract award is $4.9 million for phase one (two years) and phase two (three option years).
This prototype GIS will be the first system developed to study relationships between environmental exposures and breast cancer, and will offer researchers a new tool to conduct investigations.
The GIS layers include geographic data for general mapping as well as demographic data. Also included will be data on health care facilities, health care surveys, breast cancer and the environment.
Environmental data will cover elements such as contaminated drinking water, sources of indoor and ambient air pollution, electromagnetic fields, pesticides and other toxic chemicals, hazardous and municipal waste and radiation.
The award was made in May 1999, according to Roger Crystal, project director for the NCI program, and Cindy Wear, database engineer. The company already has build a complete GIS lab in Vienna, Va., replete with a T1 line for Internet connectivity.
Included in the laboratory's software are ArcView 3.2, a client package from Environmental Systems Research Institute Inc., based in Redlands, Calif. ESRI provides sophisticated support to users and researchers and customization via Visual Basic and Avenue, which is the ESRI scripting language.
There is also ArcInfo 8, the ESRI software supporting the GIS specialist, ArcSDE 8, ESRI middleware providing a software layer between Oracle RDBMS and GIS front-end software, such as ArcView. This allows tighter integration between ESRI GIS software and functionality and relational data, and offers the user access to a range of data rather than just the attribute information available in traditional GIS.
For database software, the company is using Oracle 8i RDBMS and Intermedia, an Oracle RDBMS extension that provides storage, manipulation and access capability for text and image data in the Oracle RDBMS.
The laboratory hardware includes a Sun 450 database server, NT Servers for hosting Web servers and NT Workstations for client software. Remote access will be through emulation software by Citrix Systems Inc., Fort Lauderdale, Fla., or a Web browser.
"We are still in the data gathering phase and are not creating new data," Crystal said. "We are working with existing databases from federal, state, local and private sources."
Among the major challenges Wear noted was collecting existing data sets from a variety of sources to build a single data set in an Oracle environment. Already, AverStar has reviewed 90 to 100 data sets, selecting the most promising and processing them. What is different about this GIS project is the creation of a data warehouse.
For Wear, a data set refers to a data source, or set of data that could be in a variety of formats. This includes a shape file, relational table or work sheets. Further, a data set can contain descriptive data (columns and values) and geographical data, or it may just be descriptive data that you traditionally find in databases such as Oracle, Microsoft Access or a flat file, such as an ASCII text format.
As in most data warehouse projects, data cleaning and transformation are a big part of the effort. Both agree that an important part of the learning curve was confronting the data compatibility issue.
"Data sets do not crosswalk easily," Crystal said.
Thus, those who intend to pursue such health-related GIS projects are well-advised to put in place a process to handle data cleaning, migration and transformation. Also, data acquisition can represent a significant hurdle, since there are many organizations from which to gather it. The difficulty is the range of policies that various local agencies have on how they will release data.