DATA INTEGRATION FROM MULTIPLE SOURCES: ENHANCING THE U.S. CENSUS BUREAU’S TIGER FILE
U.S. Census Bureau, Geography Division, Washington DC, USA
The U.S. Census Bureau’s TIGER File, a national spatial database, provided the geographic framework for the 1990 and 2000 decennial censuses. Census programs, such as the new annual national American Community Survey, require an up-to-date spatial network of features and boundaries to support on-going data collection operations as well as data tabulation functions and dissemination services. Advances in technology, such as the use of global positioning systems, require greater spatial accuracy to assure the correct assignment of counts to census areas. These types of developments require improvements and enhancements to TIGER so that geographic and cartographic services can continue to serve diverse users of this national resource.
Procedures for acquiring, processing, using, and maintaining spatial data varies among communities. Today, current spatial data oftentimes are managed more effectively in the U.S. by local, tribal, and state governments as well as private sector companies and other federal agencies for specific ventures. In addition, there are different approaches for acquiring updated spatial data. One method involves using a local file as a source to update TIGER. Information on each spatial data source, such as coverage, availability, and other metadata is maintained in a separate database to aid in determining its usefulness. Another method of data acquisition uses a current version of TIGER, covering the participant’s area, in which updates are made directly to the provided TIGER data and returned for inclusion in the national file. Other methods add to options for greater local participation such as the use of a transaction file that contains only updates and changes when compared to the original database.
Processing and ingesting various spatial data sources poses technical and operational challenges. Currently, a mix of manual processes combines with automation that includes software, standards, database technology, and related tools. Methodologies are applied based on variables such as the data source, file format, and the method in which the data were acquired. Ultimately, full automation is the goal.
Experiences in the development of the Census Bureau’s data integration objectives, operations, processes, results and lessons learned are of interest to cartographic organizations with similar spatial data missions. An exchange of these approaches with others in the cartographic community will advance further development of this work.