Census Data for Transportation Planning

Public Use Microdata Samples & Public Use Microdata Areas

Site Navigation 

Notes & News
CTPP 2000
Internet Mailing List
Old Data (pre-2000)
Urbanized Areas

Public Use Microdata Samples (PUMS)

Public Use Microdata Sample (PUMS) data is based on individual census records and is available only at large geographic areas. Think of the PUMS dataset as the raw, disaggregate census sample data, at a large enough geographic level necessary to protect the confidentiality of the census respondent. With the decennial census there are two PUMS data files:

The 1-percent PUMS is based on 1-in-100 census responses, all of which are from the census "long form." The 5-percent PUMS is based on 1-in-20 census responses, again, all of which are from the "long form." For example, in a region with 1,000,000 households, the analyst should expect about 10,000 individual household records from the 1-percent PUMS and about 50,000 individual household records from the 5-percent PUMS.

Record layout for the 1-percent PUMS file is available on the Census Bureau's web site.
There are numerous transportation planning applications for PUMS data that the transportation planner should think about, including:
  • Synthetic population microsimulation (TRANSIMS & other Models);
  • Discrete choice models (e.g., household vehicle & worker level models);
  • Weighting and expansion of household travel surveys;
  • Descriptive statistics of various markets, for example:
    • Average age of workers by means of transportation;
    • Average commute duration by age and sex of worker;
    • Workers by educational attainment by means of transportation;
    • Workers by year of entry or citizenship status by means of transportation;
    • Persons by race/ethnicity by disability status by poverty status;
    • PUMA-to-PUMA workers by means of transportation by ... by ... by ....
An exciting new aspect of the Census 2000 PUMS will be the option of analyzing the detailed characteristics of workers by PUMA of residence -to- PUMA of work, and workers at PUMA-of-work. This is different than the 1990 Census PUMS which only included county-of-work in the datasets.

Public Use Microdata Areas (PUMA)

The PUMS data is released at geographic areas called Public Use Microdata Areas, or PUMAs. The minimum size threshhold for the 5-percent PUMAs is 100,000 persons, and the minimum size threshhold for the 1-percent "super-PUMAs" is 400,000 persons.

 PUMAs were defined by the State Data Centers in cooperation with Regional Data Centers and Affiliate Data Centers. This PUMA delineation process was completed in spring 2001, and are based on Census 2000 population counts from the PL 94-171 databases. Details of the PUMA delineation process are outlined in an April 2001 letter from the Census Bureau to the SDCs (PDF file). In addition, USDOT staff prepared a guidance paper on defining PUMAs for transportation planning purposes (PDF file).

 The boundaries of the 5-percent PUMAs are included in the currently available 2001 version of the Census TIGER/Line files. Boundaries for the 1-percent PUMAs will be included in the 2002 version of TIGER/Line. For more information on PUMAs within TIGER, visit the Bureau's TIGER site at: www.census.gov/geo/www/tiger/.

 For more information on the decennial census PUMS program, visit the Census Bureau's PUMS 2000 page at: http://www.census.gov/population/www/cen2000/pums.html

Page last updated: September 24, 2003