About Census Data

The US Census of population and housing is an important resource for understanding places. The census provides information on diverse topics from population density, to rents and mortgages, to commuting modes and times -- thousands of variables that can be mapped. This document provides some background on what census data are and provides some tips for GSD students and faculty on how to get census data.

Understanding Census Data and How to Get It

Decennial Census Survey
Census Geography
Census Time Series
Getting Census Data
A Note About Normalization

The Decennial Census Survey

The US Bureau of the Census is responsible for collecting information used to dertermine allotments of governmental resources (including congressional representatives and education funding.) The constitution of the U.S. requires that in years ending in 0 each household of the united states recieves a questionaire (short form) regarding income, ethnicity and housing, every sixth household receives a long form questionaire with very detailed questions including their commute, rent, number of bathrooms, etc. Details about the decennial census can be found at the Census Web Site

Census Geography

In order to protect the confidentiality of individuals, the bureau releases only summary statistics for geographic areas: blocks, block-groups and tracts.

Population Density at three levels of geography

Census Blocks In urban areas, census blocks conform approximately to what we think of as city blocks. At this fine level of geography, the census only releases a subset of the data short-form questionaire.

Block Groups These areas are supposed to contain approximately 1200 people, but the actual count of people per block group varies widely. All of the short and long-form data is summarized at the block-group and tract level.

Tracts Tracts are larger than block groups.

A Note about Normalization

The fact that the census data are aggregated to aerial units (blocks, blockgroups and tracts) which are not consistent in terms of their geographic size or their population makes it tricky to compare one area with another without first doing a little math to normalize the units, either by dividing by the geographic area to calculate density or by factoring out the the size of the total population to determine a normalized rate. Click here for a treatice on normalization. Before spending a lot of time collecting census data, it is a good idea to work through the Mapping Census Data Tutorial so that you have an idea of what data items you will need.

Time Series with the Census

Because the census has had digital products for several decades, it is possible to estimate change over time. But it is not easy for several reasons. In 1970 there was no block or block group level data. Over the years, the boundaries of tracts have changed. The questions on the census forms have also changed.

A commercial company, Geolytics has massaged the 1970, 1980 and 1990 and a limited amount of 2000 data in a product called Census CD 40Year. This has reinterpreted 4 decades of census data to the boundaries of the 2000 census. The GSD has licenses to most of the Geolytics products. See Getting Census Data at the GSD.

For an example of a map showing time series census data, Click Here.

Getting Your Data

For many reasons it is best to get your data from the primary source. Unfortunately, the Bureau of Census makes this really difficult. There are many confusing options, each of them with particular shortcomings that make them practically impossible to use for mapping -- beginning with the fact that the geometric information about tabulation areas is only made available in an extremely arcane format known as TIGER-Line. One workflow for downloding tables of census data for mapping begins at the Census Bureau's American Factfinder. Click the Download Center) link at the left. Using this tool will get you a text file that has values delimited by a "|" (also known as a Pipe Delimited file. It will open in excel. YOu can get gis files for census geometries from Harvard's Geospatial Library Then you can join the demographic table with the geographic table using a a table join.. One annoying aspect of this workflow is that the download center does not give you the Geographic Area Charactaristics, inclusding the Land Area of each tabulation unit, which is necessary for computing density.

Easy Census through Geo/ESRI_Data

By far the easiest way to get basic census data at the GSD (if all you need is some very limited population data at the blockgroup or block level) is to use the ESRI Data compilation that can be found in L:\public\geo\esridata_9\usa\census. Even though this collection of data is relatively shallow, it is easy, it has block-level population, and automatically comes with shapefiles!

Geolytics -- the preferred method for full census data access

A reasonable person might ask why the census bureau doesn't have an easier-to-use web site. But if you compare this with the accessibility of census data in other countries, you would be glad to have any sort of access at all without paying a lot of money. Another thing we can be thankful for is that there is no copyright regulation of this public data. This means that The difficulty of getting census data on line has created a niche for private companies, such as geolytics to repackage data in a product that makes it easy to get the data you need and connect it with maps. If you are on the GSD network, you can access many Geolytics products directly from your desk. To learn about how to use geolytics at the GSD, see the manual page named Getting Census Data at the GSD Using Geolytics tools is a tad awkward, but it does work!

Downloading Census Data through Social Explorer [tm]

The Harvard Libraries have bought a subscription to Social Explorer which allows you to make census maps on line or to downl;oad tables of census data. This may be the quickest way to make a census map on-line. One advantage of social explorer is the depth of historical data that is available here. One downside is the fact that GIS shapefiles are nto available here. Another downside is that data are not available at the block level.

The National Historical GIS

The National Historical GIS at the Minnesota population center looks promising -- it seems to have all of the historic data that are in Social Explorer, and the ability to download shapefiles. Unfortunately, at the time of this writing, the Data Finder does nto seem to be working :(

The Census Transportation Planning Package

Click Here to see smoe amazing products derived from the responses to the census questions about the journey to work.