Proposal by Jon Sperling/HUD
Enabling the National Spatial Data Infrastructure (NSDI): The Need for a National Geocoding Service Center (NGSC)
Several agencies currently maintain or would like to develop enterprise geocoding service centers (e.g., HUD, EPA). This development reflects the growing interest , use, and integration of geographic information systems with existing information technology to better manage each agencies programs and assets. Addresses are the primary spatial key for visualizing, querying, and integrating most federal agency resources and assets. One of the major impediments to enabling a more robust NSDI is the poor quality of existing address information, inconsistent use of base maps and geocoding services within and across agencies, and lack of understanding within agencies of the critical importance of address information in a geospatial IT environment.
The Geospatial One-Stop Initiative provides a means to better address the growing needs to share and integrate data sets, within confidentiality constraints, across agencies. Many seemingly disparate agencies face similar problems and opportunities that can be addressed by geospatial technologies. The federal government incurs tremendous waste, duplication of effort, and inconsistent levels of data quality by the use of multiple individual strategies and systems to solve similar problems.
Geocoding or the assignment of addresses or physical locations to positions (often lat/long) on the earth's surface, is dependent on base maps, typically street level maps (or parcel maps at local level). There are no standard base maps or geocoding services within or across agencies. To some extent, this situation reflects different needs. Agencies are beginning to understand the critical importance of address data, standards, formats - not just for mailing purposes but for locating addresses on the surface of the earth and as a means for integrating data across the agency. Also, the growing use of GPS field data collection and imagery will require a more accurate national street level network.
Within this scenario, I propose the vision of a NGSC available for use by, initially, all federal agencies (in practice, this could become a model for state and local governments as well). This will ensure consistent standards across agencies and confidence in the quality of integrated federal data sets. The nation has invested heavily in the development of the Census TIGER data base over the past 40 years. TIGER was built by the active cooperation/participation of federal, state, local, private, and academic organizations. Indeed, most commercial companies have not only used this database as a primary source for building or enhancing their own national street databases but used TIGER to help model their own databases. TIGER modernization promises positional accuracy within five to ten meters with residential or business addresses receiving a GPS location. Indeed, the modernization of TIGER will require the ability to handle GPS data collection devices.
On a broader scale, it is increasingly redundant and disruptive for different federal agencies to use varying public and private versions of national street level databases. Data from separate agencies should not have to be continually re-geocoded to meet needs of each agency's spatial database. A common feature database or one with transparent links between spatial data sets will enable one-stop geocoding and passing a set of latitude/longitude coordinates for an address only once. Public and private resources should be leveraged to help build a more accurate public spatial data infrastructure. Commercial firms need to focus more efforts on value added features and attributes rather than duplicating existing public investments. Government agencies need to be more flexible and responsive to changing technology and requirements. NSDI requires a common national street centerline file or one with framework links for easy integration.
TIGER provides an appropriate initial base that will meet the needs of most agencies. Gradual enhancements to this database over time will be encouraged by federal adoption. Money saved by creating a NGSC (agency cost sharing structure), rather than multiple uncoordinated geocoding service centers in each agency, can be used to help promote and accelerate the public TIGER modernization effort.
One possibility, among others, would be to create a cross agency NGSC office within Census. Staffing needs would be minimal as the NGSC would be a highly computer intensive operation. At the same time, this office would promote and provide uniform consistent standards and technical assistance across agencies. Census may be a natural home for the NGSC for a number of reasons. Census has existing subject matter and technical expertise, ownership of a continuously maintained Master Address File (a complete geocoded inventory of all housing units and business establishments in the nation), a Congressionally mandated data sharing agreement with the USPS, and already are in the business of geocoding addresses to meet its existing mission needs. The FGDC Address Data Content Standard also was developed by the Cultural and Demographic Committee chaired by the US. Census Bureau. Another benefit of a NGSC is that taxpayers would only pay once for this service. Currently, taxpayers are paying for a national TIGER-based file and then paying commercial firms, through many agency's budgets, for "enhanced" street files and geocoding services. Most federal agencies can meet all their mapping/geocoding needs (e.g., to determine census tract for program eligibility, research, or other purpose) with the TIGER public resource.
Funding and supporting a National Geocoding Service Center offers one of the more practical and critical steps in making the Geospatial One-Stop a reality. Geocoding is one of the most important leveraging investments for NSDI promotion. Federal GIS implementation still is in the formative stages with the more traditional "GIS framework" data agencies (e.g., USGS, Census, DOT, NOAA, EPA, BLM). The next wave will include less traditional "GIS" federal agencies (e.g., BLS, DOE, HUD, DOJ, HHS). Nearly all of the traditionally non-geographic federal agencies can link and "spatialize" most of their data assets for the Geospatial One-Stop via address geocoding. Consistency and quality of this process is critical for data sharing and integration across agencies.
The OMB sponsored Geospatial One Stop Initiative, through the FGDC, provides the vehicle necessary to transform institutional relationships within the federal government as with state and local governments to support this effort. Like geographic databases in the past, geocoding systems have been built thus far to support the mandate of single institutions or parts of an institution. In order to take full advantage of the opportunities offered by new technologies, the federal government needs to develop, support and fund a National Geocoding Service Center. A further recommendation enabled by this proposal would be a more robust integration/coordination between FGDC and FedStats (e.g., National Atlas + Statistical Abstract).
Proper address management and standardization across federal data systems are critical strategies in making government more efficient and accountable. Poor address data costs billions of dollars; enables waste, fraud, and abuse; and compromises data quality, evaluation, and monitoring efforts across all programs, and inhibits the potential of data sharing and integration across agencies. Good quality address data is critical, and will become increasingly critical, in meeting OMB's strategic goals. Stakeholders or clients of government programs, including grantees, projects, etc receive grant money and mortgage guarantees, etc. based on location. Address information is critical not only for ongoing program operations but for monitoring and oversight.
Currently, most government agencies have little, if any, standardization of address data within or across programs. A Senior HUD Executive once mentioned that after a cursory review of some address files, "Albuquerque" was spelled 11 different ways. In addition to simple misspellings and typographic errors, most address data in government systems have never been verified as actually existing and thus cannot be geocoded accurately.
Most government agencies have no system or process in place to verify address data before it is entered into a database. Agencies or clients to agencies enter data, typically in a hurry, without proper understanding of its importance, and without verification. Errors and problems are often found "down the road" and cleaning it up later proves expensive and with quality concerns. This method is inaccurate, very costly, and substantially degrades data quality. Without accurate address data, there is also potential for waste, fraud, and abuse in government programs.
The solution to these problems is an enterprise approach to address management and standardization. Each agency should maintain a centralized (or distributed depending on available technologies), standardized, address database for all of its clients and data assets. All of an agency's computer systems could then tie into a common address database. Greater integration within and across programs will eliminate the current extensive amount of duplication.
OMB should support and fund enhancements that embed powerful address management and address matching, or geocoding, systems into existing data entry processes across federal programs, leading to "bad" addresses being fixed and geocoded instantly (in most cases). When a particular organization wants to do business with a federal agency, they should register their address, phone number, fax number, email address, etc. so that when they apply for a grant or other service, address data is already stored. Grantees and others using government services will be issued a User ID and password. When they log in, their address will already be displayed without need to retype it. The entered address information will then be interactively geocoded-on-the-fly to verify the geographic data (latitude, longitude, exact street address, street type [road, street, lane, boulevard, etc], ZIP code, census tract, agency region code or field office).
DRAFT 9/17/2002