Authors: Mohammad Salem*, U.S. Bureau of the Census
Topics: Applied Geography
Keywords: Census, data quality, geocoding
Session Type: Paper
Start / End Time: 3:20 PM / 5:00 PM
Room: Studio 7, Marriott, 2nd Floor
Presentation File: No File Uploaded
The U.S. Census Bureau’s Master Address File (MAF), the most comprehensive address database in the U.S., is derived from a number of sources, primarily the U.S. Postal Service’s Delivery Sequence File, and also tribal, state, and local government address information as well as Census Bureau field operations. These sources provide a large volume of original and beneficial data to assure the currency and completeness of the MAF. However, update and maintenance of a large, aggregated address database comes with numerous challenges to ensuring data quality, such as variation in source content and format, appropriate and necessary attribution, detection of duplicate addresses, and linking between multiple variants of an address. Each of these conditions, when introduced into the MAF, may adversely affect the efficiency of census and survey operations. This presentation discusses the various processes that are integral to maintaining an accurate address frame to support the decennial census and various demographic surveys.