Your behavior appears to be a little unusual. Please verify that you are not a bot.

TIGER Database Historical Perspective

October 27, 2009  - By 1 Comments

In my last two columns, I’ve made several references to geospatial data. Jon Sperling, Ph.D. GISP, wrote in and commented on the omission of the importance of TIGER data in the history of geospatial data development and commercialization. He made reference to a document he wrote that was published in 1995 regarding the development and maintenance of the TIGER database. I’ve decided to reprint his work, with his permission, as this week’s column. It gives keen insight into the early stages of TIGER. Albeit quite in-depth, it’s a fascinating read for gaining a historical perspective on geospatial data development.

Keep in mind that this document was written in 1992 so there are references to initiatives, etc. that were subsequently developed and that you enjoy today.

Sperling argues that our public investment in TIGER need not be just history but a pro-active means to leverage partnerships and new advances and innovations (e.g. synergistic links with national parcel data initiatives, local/state/federal data sharing and a national road network). Census still remains a pioneer in supporting and furthering geospatial science at all levels of our society for the betterment of our Nation’s communities.

Dr. Sperling is currently a Senior Researcher of Geographic Information and Analysis at the Office of Policy, Development and Research for the U.S. Department of Housing and Urban Development. He has designed and led many innovative geospatial, addressing, and data integration efforts in coordination with local and state governments and the academic research community. Dr. Sperling was involved in the initial development of the Nation’s TIGER/Master Address File databases for the 1990 and 2000 Census, enabling digital spatial data sharing capabilities, and efforts to enhance its coordinate accuracy and data quality. Currently, he is working with university partners on a number of innovative research applications to enable sophisticated querying of unstructured text and tables using textual spatial references in the data.

Jonathan Sperling, “Development and Maintenance of the TIGER Database: Experiences in Spatial Data Sharing at the U.S. Bureau of the Census,” in Harlan J. Onsrud and Gerard Rushton, eds., SHARING GEOGRAPHIC INFORMATION (New Brunswick, NJ: Center for Urban Policy Research). Copyright 1995 by Rutgers, The State University of New Jersey. Reprinted with permission.


Development and Maintenance of the TIGER Database: Experiences in Spatial Data Sharing at the U.S. Bureau of the Census (1992)

The U.S. Census Bureau has played, and will continue to play, a vital role in the development, maintenance, and sharing of spatial and attribute data for Geographic Information Systems (GISs) on the local, regional, national, and international levels. The Census Bureau’s development of shareable geographic data files, the GBF/DIME (Geographic Base File/Dual Independent Map Encoding) Files for the 1970 and 1980 censuses, and the TIGER (Topologically Integrated Geographic Encoding and Referencing) database for the 1990 census, have provided a major impetus to the rapid growth and diffusion of GIS technology. This chapter discusses the Census Bureau’s experiences in the spatial data sharing during these two file-building projects as well as from ongoing experiences in developing Memoranda of Understanding with federal and state agencies to update and improve the spatial and attribute data in TIGER. On the basis of these experiences, preliminary generalizations are made concerning the organizational issues that may facilitate or impede the future digital interchange of spatial data.


The entrance and persistence of the nation’s lead statistical agency as a primary producer and user of both geographic and attribute data have grown out of practical but critical concerns (Tomasi 1990). The mandate to conduct a population and housing census of the entire nation and its territories every ten years, and then disseminate the resulting information  quickly and accurately, has led to the need to devise innovative ways of improving the collection, processing, and tabulation of data.  These improvements have aimed at four goals: increasing efficiency and timeliness, improving data quality, lowering costs, and providing new products for the data user.

Over the past half century, the Census Bureau has pursued innovations and led the government and private sector in computerization, statistical sampling and interviewing techniques, data processing, quality control, and cartographic techniques, to name but a few (Anderson 1989).  The Census Bureau pioneered the first large-scale commercial user of the computer with UNIVAC 1 following the 1950 census and the development of the Film Optical Sensing Device for Input to Computers for the 1960 census.  The broad use of computer mapping by local governments and private firms was strongly influenced by the Census Bureau when it developed the GBF/DIME-Files for the 1970 and 1980 censuses and the TIGER System for the 1990 census.

In addition to being one of the nation’s largest digital geographic databases — currently sixteen gigabytes — TIGER enables the automated spatial manipulation of census data for all legal and statistical entities in the United States and its territories.  Every state and local government now has the capability to create rudimentary GIS using small-area census data, publicly available extracts of the TIGER database – TIGER/Line  files – and the appropriate hardware and software.  More than 130 private vendors currently have the capability of processing TIGER/Line files, and the number is growing.

In this respect, the development of the TIGER database may be the most important “data” file from the 1990 Census of Population and Housing.  The accessibility and widespread potential application of this innovation make it qualitatively different from earlier technological developments at the Census Bureau.  The TIGER database will be used not only internally to administer future censuses and surveys (see Marx 1986 for a basic rationale of the TIGER system) but externally to provide a major impetus to the development and sharing of integrated special information systems in the public and private sectors as well as the academic community.

A common theme in the GIS literature, generated both within and outside the Census Bureau, has been the notion that although the Census Bureau developed its GBF/DIME-Files and TIGER database to meet internal Census Bureau needs, their existence has facilitated applications well beyond the scope of census-taking.  Perhaps analogous to the many commercial spin-offs that have been generated from wartime military innovations or the nation’s space program, the decennial census operation, conducted by the government’s largest non-military assemblage of people and resources, has been responsible for a number of innovations, some already outlined, that have had a major effect on the private sector and the academic research community.

The Census Bureau’s recognition of these “consumer spin-offs” traditionally has never been well articulated or coordinated because most innovations have resulted from an internal production standpoint rather than a customer-oriented design.  Recent efforts toward building a national spatial data infrastructure and the Census Bureau’s adoption of the principles of Total Quality Management and Strategic Planning create the external and internal incentives, respectively, to stimulate a change in that design.  Paradoxically, however, the fiscal constraints of the 1990s and their impact on congressional funding, while often considered an incentive for data sharing, may also function as an impediment in the early developmental stages of new technologies.


Many local agencies have been introduced to GIS by the Census Bureau through is various geographic base-building projects over the past twenty-five years.  The development of the Address Coding Guides (ACG) for the 1970 census and the GBF/DIME-Files for the final stages of the 1970 census and for large-scale use in the 1980 census were major steps toward full automation of the Census Bureau’s geographic support programs.  Althought the ACGs provided the building blocks for the later development of the GBF/DIME-Files, they lacked geographic coordinates and topological structure (Marx 1986).

The design and development of these computer-readable files involved the active participation of federal, state, local, private and academic organizations.  The Census Use Study, a small-area data research group sponsored by the Census Bureau from 1966 to 1969, was instrumental in creating and diffusing knowledge about a system that represented map features numerically for processing by a computer to create a geographic base file (U.S. Bureau of the Census 1973).  The enhancement of this body of knowledge was of critical importance to the Census Bureau because it enabled the adoption of census-by-mail enumeration methodologies for the major urban centers of the United States.

The change in enumeration procedures was a response to the increasing costs and difficulty of conducting a traditional door-to-door canvassing of the population, a growing and increasingly urban population living within areas having mail delivery by house-number/street-name address, and technical feasibility of linking such addresses with the geographic units used for data tabulations due to improvements in computer capabilities.  For the 1980 census, the GBF/DIME-Files were enhanced to cover 287 of the nation’s largest urban centers, representing more than 60 percent of the population but less than 2 percent of the nation’s land area (Carbaugh and Marx 1990).

Data sharing during the 1970s.  The development of the ACGs and GBF/DIME-Files involved data sharing with more than 300 local planning agencies.  Sharing primarily took the form of converting analog data into a computer-readable format rather than digital exchange per se.  The long-term nature of these relationships provided the Census Bureau with intimate knowledge of the resources available to local agencies, the limitations of available data, and the willingess and ability of agencies to share data.  At the same time, local planning agencies and others, through their association with the Census Bureau, became increasingly aware of the potential computer mapping, automated address matching, and spatial data analysis to meet local needs (Sobel 1978).  These relationships played an important role in the later development of the TIGER system.

Monetary and other incentives to the local agencies played a major role in the “data sharing” development of the ACGs and GBF/DIME-Files (Silver 1977).  In many cases, local agencies completed work under contract with the Census Bureau or with funding provided by the Department of Housing and Urban Development (HUD), and Federal Highway Administration (FHWA), and other federal, state, and local planning sources.  Sensitive to the fact that the development of the GBF/DIME-Files provided a practical solution to many needs of local governments, the Census Bureau offered to provide the computer programs, processing methodology, and clerical procedures for creating  and updating the file, as well as a free copy of the completed file, to each participating local agency.  The Census Bureau also offered to do the data keying and processing if agency resources were limited.  The promise of higher quality data when these files were used for taking and tabulating the 1970 and 1980 census provided a further incentive for data sharing.

As a result of this process, the Census Bureau and local officials learned that the quality and currency of existing data used by planning agencies could not be taken for granted.  For example, many local communities relied on the knowledge of a few people for the location of dwellings, address-numbering systems often were not systematic, tax assessor sources sometimes did not meet the bureau’s quality standards, and data from different agencies often were inconsistent.  Because of these situations, the initial transition to address assignment via automated processes was difficult.

The efforts of working with so many local agencies also challenged the Census Bureau.  During the file-building projects, there was turnover in project personnel and key decision makers in the local agencies as well as at the Census Bureau.  Some “champions” of the new methods were not reelected or moved on to other jobs.  The long-term benefits of building such files were not always apparent to public administrators.  Some administrators viewed the new computer technology as a threat to their role in the agency.  Also, during the 1970s, there was a general lack of understanding (personal computers had not yet been invented) and/or trust in the new computer technology and the ever-present fear of “big brother” mainframe computers held by a few large government agencies.  All these factors tended to hamper development activities and to result in considerable variation in the quality, time expended, and funds necessary to complete the GBF/DIME-Files.

Data sharing during the 1980s.  Census Bureau funding of data collection and coding by local agencies to create the GBF/DIME-Files created precedents that would later affect the pre-1990 TIGER-building efforts.  The overall reductions in federal aid to cities and states during the 1980s provided an atmosphere of “less than cooperative initial attitudes” among many local agencies struggling with small staffs and fewer resources than they had in the 1970s.  Many agencies expected to be paid for their efforts.  These situations, and the perceived difficulties inherent in working with several hundred local agencies, contributed to the Census Bureau’s decision to do most of the map and address updates for the 1990 census in-house, albeit often based on materials supplied by local agencies.

Changes in the original GBF/DIME-File format by local agencies exacerbated some early data-sharing efforts.  Many of the larger files that had been maintained by local agencies (e.g., New York City) had been adapted and enhanced to meet local planning and administrative needs.  Fiscal difficulties at the local level, time constraints on the Census Bureau, and the lack of additional programming staff in all agencies often precluded efforts to recreate the original file structure without losing the map and address updates.  In these cases, the feature and address range update work had to be redone manually by the Census Bureau from digital plots and databases.

One of the general weaknesses of the GBF/DIME-Files was that they were limited geographically and, therefore, the ability to use the files on an ongoing basis was limited to large-scale users with on-line access to mainframe computers.  Relatively few of the original GBF/DIME-Files given to local governments were updated and maintained during the 1980’s. Even fewer of the locally updated files met Census Bureau standards for direct incorporation into the TIGER database. In practice, however, these updated GBF/DIME-Files provided the Census Bureau with its first challenge in the digital spatial data exchange at the local level.

Digitized GBF/DIME-Files, attribute-rich but of mediocre
coordinate accuracy, formed the cartographic base for 345 urban centers in the 1990 TIGER database. Absolute coordinate accuracy was not a primary concern in the development of the GBF/DIME-Files, as they were used primarily by the Census Bureau for geocoding rather than mapping purposes. Although their coordinate accuracy was well below that of the U.S. Geological Survey’s (USGS’s) Digital Line Graph (DLG) files derived from their 1:100,000-scale maps, these files represented features with all their respective feature names, address ranges, and 1980 geographic area codes in their correct relative location – sufficiently accurate for taking a census (Sobel 1986).

Although the Census Bureau would have liked to provide an enhanced cartographic quality, the deadline pressures of an upcoming decennial census forced management to abandon initial plans to align this information to the USGS’s DLG files. Also, because of staff and time constraints, the files were sent to four private-sector contractors for digitizing of feature updates using Census Bureau-supplied updated 1980 census maps. The results from these arrangements were of mixed quality.

TIGER Database Development, 1983-1990

The institutional knowledge and experience gained from the development of the GBF/DIME-Files, further theoretical and conceptual advances in the field of mathematics (Corbett 1979 and White 1984), and the “enabling” availability of new and affordable technology, provided fertile ground for the Census Bureau’s next challenge: the development of the TIGER database for use in the 1990 census. Whereas the GBF/DIME-Files covered small noncontiguous portions of the United States and were developed initially without spatial or geographic references in their design, the TIGER database covered the entire nation and its territories and was grounded in a more rigorous conceptual model of topology and space (see Boundriault 1987; Kinnear 1987; Broome and Meixler 1990).

Census Bureau/USGS cooperative agreement. In addition to data-sharing arrangements on the local level, the ability to complete the building of the TIGER database in time for the 1990 census was directly dependent on a landmark 1983 data-sharing agreement with the USGS. The USGS provided the Census Bureau with computer files of scanned versions of its 1:100,000-scale maps for the lower forty-eight states. In return, the Census Bureau assigned cartographic classification codes to the roads in these files. The resultant product formed the cartographic base for all areas outside the large urban centers covered by the GBF/DIME-Files, thereby enabling the Census Bureau to complete a coast-to-coast digital map base in time for the 1990 census (McKenzie and LaMacchia 1987).

Interagency cooperation with the USGS was an experience that provided valuable lessons to the Census Bureau in particular, and a model for future cooperation between federal agencies in general. Following a successful Florida pilot project, high-level management in both agencies perceived that cooperation would result in a win-win situation. Each agency would be able to accelerate its individual map production programs and, in the process, they could develop the first “large-scale” digital map file of the United States. Early negotiations ensured that neither agency would feel it was bearing an unfair burden. A schedule of meetings on a regular basis ensured communications during all phases of the cooperative agreement. The challenge of meeting the Census Bureau’s decennial deadlines provided an added measure of incentive to “get the job done” and an ongoing requirement to measure progress.

The U.S. Census Bureau/USGS cooperative agreement minimized duplication of effort in federal map automation activities and provided immediate short-term benefits to both agencies. The success of this cooperative venture demonstrated to the Department of Commerce, the Federal Office of Management and Budget (OMB), and the Congress that there were significant benefits to be derived from such activities. The success of this cooperative effort also resonated in the later development of the Federal Geographic Data Committee (FGDC) (OMB 1990), the growing impetus for further cooperative efforts in spatial data sharing, and the increasingly articulated vision of a national spatial data infrastructure (Marx 1992).

Other success factors. The success of the TIGER System and the ability of the Census Bureau to overcome organizational inertia both within and outside the Census Bureau were due to a number of factors. There was a shared sense both within the Census Bureau and by the data-user public that change was needed in the geographic support process. A primary incentive for developing the TIGER database was the large number of inconsistencies between the statistical and geographic data products in the 1980 and earlier censuses, a product of the complex and clerically intensive preparation of maps, ACGs and GBF/DIME-Files, and geographic reference files (Marx 1986).

The enormous political and economic ramifications of the decennial census made everyone a stakeholder in the process. The growing importance of the information sector of the economy and growing public demands for more accurate, cost-efficient, timely, and accessible data products helped to promote an environment receptive to the exchange of data, expertise, and experience with other governmental agencies at all levels, the private sector, and the academic community.

Bureaucratic inertia was further overcome by staff commitment, expertise, and initiative with the Census Bureau’s Geography Division. The transition from traditional to automated mapping for the 1990 census required changes in the organization and planning of the Census Bureau’s mapping activities as well as in the requirements of staff in developing and incorporating new cartographic techniques and computer skills. Motivated by a decennial environment of schedules and fixed deadlines with no alternative to full automation for product delivery and a cooperative agreement with the USGS, the Census Bureau’s staff produced significant results in a relatively short time (Trainor 1990). The resources and skills gained from this experience, combined with a history of successful technical innovations at the Census Bureau, promoted a willingness to reach out and explore the potential for digital data sharing.

Some criticisms. Countering these positive aspects are criticisms of the TIGER database, such as the relative poor coordinate accuracy of the roads in the major urban centers (the roads that came from the GBF/DIME-files used in lieu of USGS DLG file), lack of address range and ZIP Code improvements or expansion beyond the 345 GBF/DIME-File areas, and inconsistencies in the names and classifications of streets. Public complaints about the quality of the data in the TIGER database provided the Census Bureau and its parent agency, the Department of Commerce, with further verification of the wide applicability and importance of this database beyond the internal needs of the Census Bureau.

Although valid, most of these situations were not critical for the taking of the 1990 census, the primary mission of the Census Bureau. In fact, many private consulting firms have taken advantage of these “problems” to repackage “new and improved” versions of the bureau’s publicly available extracts from the TIGER database: the TIGER/Line files. The Census Bureau is correcting many of these situations and will release future TIGER extract products with these updates and corrections.

In general, the TIGER database continues the GBD/DIME-File tradition of being attribute-rich and current but with limited coordinate accuracy in the major urban centers. The USGS’s DLG files, on the other hand, have high “ground truth” accuracy for the features they show but have few attributes (DLG-Enhanced Files, once released, will improve on the latter). In addition, the DLG Files do not contain current information and the USGS has not been provided with the financial resources need to perform frequent, nationwide, and systematic updates. As stated previously, alternative methodologies for updating the files are being evaluated, including an initiative to accelerate the collection of base cartographic data using graphic or digital orthophoto quadrangles or aerial photography (FGDC 1992).


Since the completion of the TIGER database for the 1990 census, the Census Bureau has become increasingly aware of its vast potential as well as its current weaknesses. If viewed as a process rather than a product, the TIGER database provides an opportunity to improve statistical accuracy and data quality significantly. The updated address and geographic information systems in a growing number of public and private agencies and the databases of the U.S Postal Service (USPS) provide important means of enhancing the collection, processing, and tabulation of census data.

The planned release of a new TIGER/Line extract by the Census Bureau containing extended address ranges and ZIP Codes for all areas with city-style mail delivery is significant, not only for the inherent value of the added data (there will be an increase in address range coverage from approximately 55 percent to 85 percent of all housing units in the United States), but because it represents new capabilities that never existed before. For the first time, the Census Bureau will be releasing value-added files for data users after the census that are not directly tied to the decennial statistical data products (however, they will be used for geocoding establishments in the intervening economic census). This precedent also is important in that it reflects a subtle change in the traditional once-a-decade data dissemination paradigm and opens new possibilities for future digital data exchanges. Perhaps, it also is indicative of the more customer-oriented approach mentioned earlier in this chapter.

Because the Census Bureau was not intended to be the nation’s preeminent mapping agency, any data-sharing agreements to improve the TIGER database must be perceived by the overall organization as primarily benefitting the census-taking process. At this stage of planning for the 1992 and 1997 economic censuses and the 2000 decennial census, the Census Bureau has made an organizational commitment to the integration of the Census Bureau’s related Address Control File with the intercensal update and improvement of the TIGER database. This commitment, however, is dependent on the availability of resources.

In the absence of a national updated map system, the U.S. Census Bureau has pursued a relatively high-cost mapping compilation strategy for the 1990 and previous censuses (Rhind 1991). Labor-intensive comparisons of reference sources, often of varying scale and quality, have been the primary means for updating census maps. In this respect, the Census Bureau has pursued and maintained close contacts and relationships with other federal agencies, state, regional, and local planning and transportation agencies, engineering firms, aerial survey companies, tax departments, utility firms, and a host of other public and private firms with current map and address reference source materials.

The development of the TIGER System and other digital spatial and attribute databases, coupled with the proliferation of more powerful computer hardware and GIS software, allows data exchange to occur in a more sophisticated, more timely, and potentially more accurate and less costly manner. From a technical perspective, there appear to be few limits to the advance of this new mode of digital data exchange. According to Cooke (1995), the technical problems of data sharing have mostly been, or are in the process of being, solved. The non-technical components of data interchange may prove more daunting.

Current Data-Sharing Plans

Current data-sharing plans at the Census Bureau are twofold. On the one hand, the Census Bureau is investigating the possibility of national sources, predominantly federal agencies, that could provide the information to keep the feature and address-range information in the TIGER database up to date. Based on a report by the Government Accounting Office (GAO 1991), which stated that federal agencies increased their planned expenditures on GIS by about 60 percent between fiscal years 1990 and 1992, the likelihood of such exchanges is promising.

At the national level, the Census Bureau has entered into or proposed data-sharing agreements with several large governmental or quasi-governmental agencies including the USGS, USPS, the Environmental Protection Agency (EPA), the Soil Conservation Service, the Federal Railroad Administration, the U.S. Army Corps. of Engineers, and the Federal Emergency Management Agency (FEMA). These efforts to improve the spatial and attribute data in the TIGER database are ongoing and likely will include several other agencies by mid-decade. As the coordination of GIS activities improves in the federal sector, bilateral agreements between agencies to improve the TIGER database will likely evolve into multi-agency agreements, as has already occurred at the state and local levels (see Murakami and Greenleaf 1992).

The Census Bureau also is investigating possible mechanisms for the electronic interchange of updated geographic information with state, local, private, and academic organizations. The current preference of the Census Bureau is to coordinate these efforts and control quality at the state level rather than having to deal with conflicting data and different file formats from several thousand local governments, private agencies, universities, and other sources. Experience indicates, however, that this will not always be possible.

Role of the FGDC in Data Sharing

Increased GIS use by federal agencies as well as state and local agencies has led to renewed efforts to coordinate development, sharing, and dissemination of spatial data, primarily through the Federal Geographic Data Committee (FGDC). The FGDC, formed in late 1990 at the direction of the OMB in its Circular A-16 (OMB 1990), includes representatives of fourteen departments and independent agencies but has no direct authority, responsibility, or resources. Participation is voluntary, with decisions based on consensual agreement among its members. Given the current limits on availability of fiscal resources, several agencies have agreed to pursue cooperation in accelerating the 1:12,000-scale digital orthophoto program and the 1:24,000-scale digital quadrangle program. The Census Bureau is working on individual Memoranda of Understanding (MOU) with other federal agencies in the spirit of the FGDC and has agreed to cooperate with the USGS in devising a still more powerful data structure.

In the long run, the FGDC may provide the vehicle necessary to transform institutional relationships within the federal government as well as with state and local governments, the private sector, and the academic community. The FGDC is making an increased effort to involve the non-federal community in its coordination work. Although the goals are lofty and the potential long-term benefits extraordinary, effective leadership at the highest levels and a concomitant commitment to the development of compatible standards will be necessary. In the absence of these developments, only significant short-term and real cost-saving benefits gained from data sharing by particular agencies will move the process forward toward a national digital spatial database.

Significantly, in July 1992 the Census Bureau and the USGS signed an amendment to the original 1981 Memorandum of Understanding that commits both agencies to merging the current DLG and TIGER databases, including information resulting from partnerships with other agencies. The development of a shared database that combines the essential geographic information needed by both agencies to carry out their respective institutional mandates will have a profound effect in further stimulating the development of a national spatial data infrastructure in the United States.

Intercensal Data-Sharing Projects, Post-1990

Current data-sharing experiences at the Census Bureau have been exploratory, and the actual mechanisms and standards for digital data interchange are still in the process of being developed.  The Census Bureau is engaged in a number of activities to help promote digital data sharing during the 1990s.  These activities include conferences, participation in the FGDC initiative, MOUs with other federal, state, and local agencies, pilot projects, and the planning and imminent release of the prototype version of the TIGER/SDTS (Spatial Data Transfer Standard) file.

Based on these forums and activities, the Census Bureau has begun to receive initial feedback on the non-technical impediments to and incentives for digital spatial data interchange.  The Census Bureau has provided its digital geographic and statistical data sets to the public at the cost of dissemination and as a public resource (OMB 1992).  Other public and private agencies, however, have different perspectives and regulations pertaining to their own data sets and on the updates they may perform.  Profit is a major concern of private companies, and public agencies also are looking to their products as a source of revenue.

States and local agencies consistently have reported one or more of the following problems in trying to coordinate a GIS:

•    Agencies wanting proprietary control of internal data

•    Lack of resources in one department or institution affecting data requirements of another

•    Archaic systems

•    Managers and commissioner-level officials who know, or care, little about GIS

•    Staff turnover

•    Lack of commitment

Once a state or local GIS is operational, the lack of overt incentives to expend the additional time and expense required to feed local updates into the TIGER/Line or SDTS formats for interchange become more apparent.  Ensuring the currency, accuracy, and quality of the TIGER database is an integral part of ensuring the accuracy and quality of the associated census data, which are used for reapportionment, redistricting, the distribution of federal funds, and innumerable planning and development programs – but this may be a long-term and nebulous consideration for some agencies.  The lack of a clearly defined formal process for data interchange by the Census Bureau could provide an even greater impediment to institutions willing to share data.

The existence of multiple geographic data file formats also have inhibited the interchange of data.  Future geographic data files at the Census Bureau will be released in accordance with a recent Federal Information Processing Standard (FIPS) – the SDTS.  Adoption of the SDTS involved cooperation of federal, state, and local officials, the academy community, and private sector over an extended period of time.  The Census Bureau was an active participant in the development of this new federal standard for data exchange and, to that end, released its first prototype TIGER/SDTS file (Davis et al. 1992).  However, the effect of this rather complex format on data interchange is not yet clear.

Each data-sharing agreement brings forth new possibilities, new arrangements between agencies, and the potential for new products.  Working with a variety of agencies and soliciting recommendations for improving TIGER, Census Bureau staff have noted similarities among the interests of many agencies.  For instance, conversations with the USPS and the U.S. Department of Transportation have revealed similar interests in enhancing attributes for streets (e.g., turn and directional restrictions).  In many cases, the Census Bureau would be able to use the enhancements needed by other agencies for improving the quality and cost-effectiveness of its own internal operations (e.g., routing of enumerators).

Proposed Census/USPS cooperative program.  The USPS and the Census Bureau have been working together for more than thirty years in the delivery and return of questionnaires for the decennial, agriculture, and the economic censuses and surveys.  The increasing use of mail-out/mail-back procedures throughout the nation has made the USPS an indispensable partner in the Census Bureau’s data-collection activities.  The Census Bureau traditionally has paid the USPS to verify the completeness and accuracy of its decennial census address list, which was purchased from private vendors and enhanced through in-house programs, prior to the mailout of questionnaires.  The Census Bureau also worked with the USPS on the development of computer algorithms to match the ZIP+4 files to GBF/DIME-File records during the 1980s.  ZIP+4 files, also known as the Address Management System (AMS) Files, contain potential address ranges for all areas where the USPS delivers mail.

In an effort to enhance this cooperation to the benefit of both agencies, the Census Bureau proposed a formal Memorandum of Understanding (MOU) with the USPS similar to the one it had with USGS in the 1980s. In 1990, the Census Bureau and the USPS, with the participation of the USGS, began cooperating on a pilot project to provide a better geographic database for all three agencies.  If signed, this MOU will have broad implications, not only for these three agencies, but for the GIS community as a whole during the 1990s.

The Census Bureau’s proposal envisions a four-year file update/enhancement plan (fiscal years 1994-1997) and subsequent ongoing cooperative efforts to update a Post/TIGER database.  As with the earlier USGS agreement, the USPS/Census Bureau MOU would formalize a seemingly well-suited alliance between two agencies.  Cooperative database-building activities could reduce duplication of efforts, thereby reducing overall costs, as well as improve the geographic and attribute accuracy of the information available to each agency.

The Census Bureau’s objectives for the proposed joint program during the intercensal years are (1) to obtain updated information on the location of streets with their names, address ranges, and ZIP codes as well as the location of group quarters, office building or other locations of economic activity, and (2) to improve the error-prone decennial census address lists development operations.  According to a recent GAO report (1992), enumerator follow-up to vacant and nonexistent units alone resulted in an added expenditure of approximately $317 million to the 1990 census operation.   Accomplishing the above objectives will enable the Census Bureau to reduce its critical dependence on a large, temporary clerical workforce before and during each decennial census, thereby reducing costs, and to improve the overall quality and consistency of decennial census data products.

In the overall proposal, the joint venture envisions updating the TIGER database to permit automated analysis of carrier routes and the production of carrier route maps for use by the USPS.  In the pilot study, the effort to improve and maintain the positional accuracy and completeness of the TIGER database involved the use of Global Positioning System (GPS) technology with receivers mounted on USPS-supplied vehicles.  In order to improve and update the geocoding capability of the TIGER system, the pilot study also updated the address ranges in the TIGER database and added ZIP+4 Codes and other USPS information.  The joint venture envisioned would extend similar geocoding improvements to all parts of the United States with city-style address systems.  The Census Bureau would provide the USPS with its technical and geographic expertise.

Potential impediments.  Although this data-sharing scenario appears to satisfy the needs of both agencies, there are a number of organizational, behavioral, and institutional impediments that will need to be overcome.  Some of these impediments are specific to these two agencies while other are generic to any data-sharing milieu.  One specific impediment is that the USPS is not part of the FGDC.  The USPS also is a quasi-federal agency, which means that it must justify its participation in an agreement on a benefit/cost basis including expected revenue or savings from potential products.  There also are questions about what information will be shared and what will be the property of each organization.

A more general institutional impediment is the difficult of building a single database to serve the needs of different agencies.  Federal agencies have worked independently of each other for a long period of time collecting and structuring their data according to geographic units based on their own unique criteria and naming conventions.  Data sharing presupposes a strong, long-term, funded commitment to reconcile what really constitutes different versions of the same reality.  In order to reach agreement, a shared database initially may involve compromises that could make it less effective than two separate databases.  Each agency has its own self-interest and mission that must be met first.  Short-term objectives may become more pronounced and inhibit progress.  The need to understand each other’s terminology, organizational structure, and needs represents another potential impediment.

Another important issue, not only for the Census Bureau but for all potential data partners, is the potentially differing perception of the benefits from data sharing.  One agency may believe that it is giving up more than the other agency and is bearing an unfair burden, causing a negative effect on data-sharing plans.  Some groups within each agency may either be opposed to or less than committed to the idea of data sharing.  While this issue was always in the background during the joint Census Bureau/USGS Cooperative Mapping Project, it never came to the forefront as a major issue.

Another potentially critical impediment common to bureaucracies is the inertia effect; that is, it is easier to stick with the tried and reasonably true methods than to try something new.  Change can be intimidating and can upset a delicate balance of power in an organization.


In a broad sense, the Census Bureau traditionally has been in the business of sharing geographic data and has a long history of partnerships with state and local governments.  The Census Bureau receives updated governmental unit boundaries from local officials on a periodic basis through its Boundary and Annexation Survey.  The Census Bureau also works with local Census Statistical Area Committees, composed of representatives of the public, private, and academic communities, in delineating the boundaries of census statistical areas such as census tracts, block groups, and census-designated places.

The Census Bureau also has worked closely with state election officials to provide the information they need for redistricting and reapportionment; with the U.S. Department of Education to incorporate the boundaries of school districts as a means to produce data for school districts; and with metropolitan planning organizations to improve the quality of the Census Bureau’s address reference files, which improves the quality of the usefulness of the census journey-to-work and place-of-work data, and to define traffic analysis zones in terms of census blocks to facilitate the tabulation of decennial census data for those areas.  The Census Bureau has built its massive geographic database primarily from source materials acquired from state and local agencies.  In a new program based on 1990 census data, data users may independently aggregate census blocks to define their own statistical areas and receive maps and data profiles of these user-defined areas from the Census Bureau on a cost-reimbursable basis.

In addition, the Census Bureau maintains and is constantly improving its huge, ongoing institutional data-sharing apparatus.  Each of the Census Bureau’s twelve regional offices supports information services and geographic programs that coordinate activities with state data centers and their affiliates throughout the United States.  The Census Bureau’s Data User Services Division combines educational functions (e.g., ongoing workshops on TIGER, census maps, and data products) with its data distribution functions.  The Census Bureau participates in the kindergarten through twelfth grade (K-12) geographic literacy campaign in the United States a means to address the important issues of access to new information technologies.  Public access issues (Emergency Planning and Community Right-to-Know Act 1986) also have provided the impetus for an ongoing cooperative project with the EPA and the National Oceanic and Atmospheric Administration (NOAA) that supports the development of public domain software that links and displays environmental, socioeconomic, and demographic data using an extract of the TIGER database.

From a more global perspective, the Census Bureau shares its technological and statistical know-how with a number of participating countries through its in-house and overseas training programs.  The Census Bureau also is cooperating with Statistics Canada (Haythornthwaite 1992) and pursuing talks with Mexico’s statistical agency, Instituto Nacional de Estadistica, Geografia e Informatica, to create a North American Common Borders Database.  In summary, the Census Bureau maintains a huge ongoing apparatus that supports and complements its current efforts at spatial data interchange.


As the Census Bureau prepares for the 2000 census, change, once again, appears imminent.  According to a recent GAO report (1992,4), “the current approach to taking the census appears to have exhausted its potential for counting the population cost-effectively.”  Similar statements were made in reference to the Census Bureau’s geographic support program following the 1980 census (Tomasi 1990).  These statements have recurred with periodic frequency in the recent history of census-taking.  In large part, the innovations that followed were a response to intense public scrutiny, a result of a process that has significant and far-reaching political (reapportionment and redistricting) and economic (distribution of government funds) consequences.  Over the past fifty years, the Census Bureau has met not only the challenges brought about by massive social, demographic, and economic change, but has provided effective leadership in applying and diffusing a number of new technologies to both the governmental and non-governmental sectors.

The application and use of new technologies on a vast scale by public organizations are reshaping the internal organization and relationships within the public sector as well as among the public and private sectors and the academic and research communities.  The dominant values of narrow, functionally separate governmental agencies and departments are increasingly being replaced by a system of greater complexity and interrelatedness and a growing trend toward public and private sector cooperation.  The Census Bureau’s TIGER system is an important example of this phenomenon and may be a primary motivator for such changes.

The Census Bureau’s geographic and statistical products – inexpensive and ubiquitous spatial and attribute raw material – have helped transform GIS from a highly technical field dominated by large agencies, private firms, and universities, to one that is becoming increasingly accessible to many data users.  Advances in our technological infrastructure thus far have been the primary factor enabling institutions to develop and share digital geographic data.  Personal computers, workstations, CD-ROMS, and databases available for automation were relatively nonexistent until the 1980s.  The incipient use of new telecommunication technologies, such as the Internet, may have an equally profound effect on data sharing and online services in the latter part of this decade.

As computer hardware and software continue to become less expensive and more powerful, the vast potential of GIS will depend increasingly on the behavioral, organization, and institutional issues acting as impediments and incentives to the sharing of geographic data.  The integration of various multi-media technologies and the growing capability to link a wide variety of public and private databases also raise a number of privacy issues.  Deriving the full benefits of GIS and related information systems will depend, to a significant degree, on how society approaches and resolves these issues (Onsrud 1992).

Based on past experience, it will require extraordinary leadership, communication and flexibility among agencies to facilitate the process of spatial data interchange.  Data sharing will be most successful when such ventures can be justified by short-term results, verified cost reductions, improved operations, and minimal problems of data ownership.  Data sharing will be enhanced to the degree that these ventures can be built around joint development projects such as the Census Bureau/USPS and the Census Bureau/USGS experiences in the 1990s.  Awareness of the need for such coordination is growing and is best reflected in the increasing number of statewide GIS committees and partnerships between the federal and state communities.

Geographic databases have been built thus far to support the mandates of single institutions or parts of an institution.  All who collect and manage data for activities related to their own responsibilities will need to understand and appreciate the value of those data to others and to collect and structure their data accordingly.  In order to take full advantage of the opportunities offered to these new technologies, business, government, and academic will need to develop, support, and fund data exchange on a systematic and ongoing basis as well as promote accessibility of GIS capabilities to all sectors of our society.



Anderson, M.J. 1988. The American census: a social history. New Haven: Yale University Press.

Boudriault, G. 1987. Topology in the TIGER file. Eighth International Symposium on Computer-Assisted Cartography, Proceedings. Baltimore, Maryland, 258-263.

Broome, F.R. and D.B. Meixler. 1990. The TIGER database structure. Cartography and Geographic Information Systems 17, 1:39-47.

Carbaugh, L.W. and R.W. Marx. 1990. The TIGER system: a Census Bureau innovation serving data analysts. Government Information Quarterly 7, 3:285-306.

Cooke, D. 1995. Sharing street centerline spatial databases. In H.J. Onsrud and G. Rushton, eds., Sharing Geographic Information. New Brunswick, NJ: Center for Urban Policy Research, Rutgers University.

Corbett, J.P. 1979. Topological principles in cartography. Technical Paper 48. U.S. Bureau of Census, Washington, D.C.

Davis, B.A., J.R. George, and R. W. Marx. 1992. TIGER/SDTS: standardizing an innovation. Cartography and Geographic Information Systems 19, 5:321-327.

Emergency Planning and Community Right-To-Know Act: Title 3 of the Superfund Amendments and Reauthorization Act of 1986. PL 99-499, 17 October 1986. United States Statuates at Large 100. pp. 1728-1758.

Federal Geographic Data Committee. 1992. Multi-agency initiative to meet high priority requirements for base cartographic data. FGDC Subcommittee of Base Cartographic Data. Washington, D.C.

Government Accounting Office. 1991. Geographic Information Systems: information on federal use and coordination. IMTEC 91-72-FS. Washington, D.C.

________. 1992. Decennial Census: 1990 results show need for fundamental reform. GAO/GGD-92-94. Washington, D.C.

Haythornwaite, T. 1992. Development of the United States-Canada Common Border Database. The Operational Geographer 10, 1:28-30.

Kinnear, C. 1987. The TIGER Structure. Eighth International Symposium on Computer-Assisted Cartography, Proceedings. Baltimore, Maryland, 249-257.

Marx, R.W. 1986. The TIGER System: automating the geographic structure of the United States census. Government Publications Review 13, 181-201.

________. 1992. Building the National Spatial Data Infrastructure: the data integrity challenge. Paper presented at the Regional Surveying Engineering Conference, Hartford, Connecticut.

McKenzie, B.Y., and R.A. LaMacchia. 1987. The U.S. Geological Survey-U.S. Bureau of Census Cooperative Digital Mapping Project: a unique success story. Paper presented at American Congress on Surveying and Mapping meeting, Reno, Nevada. Fall.

Murakami, E., and K. Greenleaf. 1992. Multi-agency TIGER file updating. URISA Proceedings 2:25-35.

Office of Management and Budget. 1990. Coordination of surveying, mapping, and related spatial data activities. OMB Circular A-16 (Revised). Washington, D.C.

_______. 1992. Management of federal information resources. OMB Circular A-130. Washington, D.C.

Onsrud, H. 1992. Privacy and spatial databases. Technical Program Abstracts., 27th International Geographical Congress, Washington, D.C., 480-481.

Rhind, D.W. 1991. Counting the people: the role of GIS. In D.J. Maguire, M.F. Goodchild, and D.W. Rhind, eds., Geographic information systems: principles and applications. Longman Scientific and Technical, Essex, 2:127-137.

Silver, J. 1977. The GBF/DIME system: development, design and use. Paper presented at 1977 Joint Annual Meeting of the American Society of Photogrammetry and the American Congress on Surveying and Mapping. U.S. Government Printing Office, 1977-240-869/1102.

Sobel, J. 1978. GBF/DIME system – development and reference source problems. Applied Geography Conference, SUNY – University Center at Binghamton, 1:112:121.

_______. 1986. Principal components of the Census Bureau’s TIGER file. Research in contemporary and applied geography: a discussion series. SUNY at Binghamton, 10, 3:1-17.

Tomasi, S.G. 1990. Why the nation needs a TIGER system. Cartography and Geographic Information Systems 17, 1:21-26.

Trainor, T.F. 1990. Fully automated cartography: a major transition at the Census Bureau. Cartography and Geographic Information Systems 17, 1:27-28.

U.S. Bureau of Census. 1973. Census Use Study. International DIME Colloquium. Confernce Proceedings, Washington, D.C., August 27-29, 1972.

White, M. 1984. Technical requirements and standards for a multipurpose geographic data system. The American Cartographer 11, 1:15-26.

This article is tagged with and posted in GSS Monthly
Eric Gakstatter

About the Author:

Eric Gakstatter has been involved in the GPS/GNSS industry for more than 20 years. For 10 years, he held several product management positions in the GPS/GNSS industry, managing the development of several medium- and high-precision GNSS products along with associated data-collection and post-processing software. Since 2000, he's been a power user of GPS/GNSS technology as well as a consultant with capital management companies; federal, state and local government agencies; and private companies on the application and/or development of GPS technology. Since 2006, he's been a contributing editor to GPS World magazine and the Geospatial Solutions website. He is the editor of Geospatial Solutions Monthly, a weekly newsletter focused on geospatial technologies. Follow Eric on Twitter at @GPSGIS_Eric.

1 Comment on "TIGER Database Historical Perspective"

Trackback | Comments RSS Feed

Post a Comment