Overview of the SoftwareThis guide is primarily concerned with using US Census data with ArcGIS, although it does touch on some other types of data.
At least take a look at the above before collecting your data and continuing.
COLLECTING CENSUS DATA
Create a place to save your files! Don't just dump them anywhere on the hard drive or you might lose them. The "D:" segment of the hard drive of computers in Room 226 doesn't get erased, or use a USB thumb drive or portable hard drive, or your AFS space.
FIPS County Codes
Most Census data use filenames based on the Federal Information Processing Standards (FIPS) Codes as published by the National Institute of Standards and Technology. You can refer to the county codes at the NIST website.
These geographic data (census boundaries, roads, rivers, etc.) are available for download to members of the Duke community (NetID required).
- Geographic Layers
The Census TIGER/Line files contain spatial information such as census and political boundaries. ESRI has converted them from the TIGER format into Shapefile format, which can be read in ArcGIS.
MSA's: The Office of Management and Budget (OMB) defines Metropolitan Statistical Areas (MSAs) as a group of counties that encompass the metro area. To determine which counties make up a metro area during a particular census, the Census web site has MSA listings showing the metro areas that were in effect for past decennial censuses back to 1950. (They also have the most recent (post-2000) lists of MSAs and related areas at the Census web site, along with policy documents relating to the definitions.)
Note: In 2000, the MSA's for large metro areas were divided into Primary Metropolitan Statistical Areas (e.g., Dallas) and Consolidated Metropolitan Statistical Areas (e.g., Dallas-Fort Worth). Check with your instructor regarding what level to use for an assignment. Since 2000, the OMB defined Metropolitan Divisions as subcomponents of some Metropolitan Statistical Areas. MSA's can be grouped into larger aggregations called Combined Statistical Areas (roughly equivalent to the old Consolidated MSA). Currently, the Raleigh-Durham-Cary CSA consists of the Raliegh-Cary MSA and the Durham MSA along with the Dunn Micropolitan Statistical Area.
ESRI also makes available the Census Demographics files (2000 data) for various Census geographies, which you can include in your download along with some geographic layers. They have a geographic identity field with values that match values in the attribute fields of the geographic layers for easy joining to GIS layers.
- The SF 1 Quick Reference Guide lists the field names for fields in the attribute tables and in the demographic files. This is especially important for the demographic characteristics files, since there are so many and the field names can be terse. The PL 94-171 Quick Reference Guide provides the field names for the vast number of permutations of answers to the question regarding racial identity in the 2000 Census.
- Unfortunately, ESRI does not provide social and economic data (from the SF3 files, or sample data) for download. You will need to get these from other sources.
What You Download
- Your file will be a single zip file containing, in turn, zip files for each layer you download (a single layer may consist of several files). You will need to click "Open," which will download the file into the WinZip program. (The library computers won't let a user download a file directly to the hard drive, but will let WinZip save a file to it.) The program Win-Zip is on the Library computers to unzip the files.
- Key to Filenames: Included in the zip file is a key to the filenames of the TIGER layers, to help you remember what you downloaded.
- The GDT Major Roads layer can be especially useful layer for context (compared to the overly busy TIGER streets layer), but these cover entire states, not just counties. The attribute tables of some layers (such as census tracts) already contain many fields of data. These data are also available on CD's or DVD's. Ask the GIS or Data Services Librarians for assistance.
Includes tract level data from 1970 to 2000 on one disc, both from the short form 100% data (SF1) and the long form sample data (SF3). These are a convenient way to select variables and geographies of your choice and export the data into dBase files or geographic boundary files as ArcView Shapefiles (which will also include attributes in dBase format).
Besides the online database, this is on a CD-ROM. See our Step-by-Step Help Guide for procedures on using data from these CDs with ArcGIS.
The NCDB might be the easiest method for incorporating tract-level social statistics (e.g., income and poverty) into a project, especially if you need need to get data from several states (e.g., you can choose MSA as your level of geography and get one file with all tracts in that MSA for the chosen year).
Besides our step-by-step help guide, a tutorial on the process using an older Geolytics CD has also been done by the Electronic Data Center of the Columbia University Libraries. The interface is slightly different with the NCDB product.
Tract Level Only: Remember, that to get 2000 boundary files or data for geographic entities besides tracts, you'll need to get the files from ESRI or elsewhere. You will probably want some other layers for context (e.g,. county boundaries, water features, highways, railroads, etc.).
(U.S. Census Bureau)
To get demographic data directly form the source, see our separate help guide.
Land Use - Land Cover data (LULC)
WebGIS is a site that provides free LULC data already converted to ArcView shapefile format. Unprojected Latitude/Longitude coordinates will work best with Census data. Most of the data created by or provided by the US Geological Survey or EPA is in formats that aren't ready-to-use in ArcGIS.
Each 1:100,000 scale LULC quadrangle covers 2º of longitude E-W and 1º of latitude N-S. WebGIS provides an index graphic that allows you to click on the counties or to choose from an alphabetical list. It will then list all 1:100,000 quadrangles that cover at least part of that county.
You will need to join a dBase file to the downloaded LULC layer's attribute table in order to match English words to the numeric LULC classification codes. (Field LUCODE in the attribute table matches to LU2-CODE in the dBase file.)
These pages lead to many other GIS data sources. You may have trouble obtaining geospatial data at the necessary geographies or in a format that is immediately usable in ArcGIS. For instance, crime data might only be available for an entire city, county, or MSA. If you need it broken down by police precinct or if you need a point file showing locations of crimes, you may need to contact the municipality. Also, when the geographies don't mesh with Census geographies, you will need to obtain or create boundary files to match the data.
GIS Links: More downloadable data in various formats.
- An online guide goes over some of the several CD-ROM sources for geospatial data in ArcView format. WARNING: it hasn't been updated in a while, but does mention some sources for 1990 Census data and for earth science data.