All About ZIP Codes: 2010 Supplement
Supplement to What?
In the late 1990s we created the original ZIP Code Resources Page, containing a variety of information and tools related to ZIP codes. While much of what we talked about in that document was not time-specific, some of it was. It talked about the status of ZIP codes in the mid-90's and links to data all refrenced then-current 1990's-vintage resources. Around 2001 we created the updated Version 2 of the document. References to various datasets and tools were updated to point to the new post-2000 vintage resources. And now (2012 - 2013), the 2010 census has come and gone and we are dealing with the 2010 cycle of ZCTA's from the Census Bureau, plus a new way of getting demographic data involving the replacing of Summary File 3 (decennial census product) with data based on the new American Community Survey (ACS). Rather than create an entire new Version 3 document we have decided, at least for now, to merely create this supplement (to Version 2). Here we shall deal only with what's new in the world of ZIP codes in the decade of the 2010's. For general information, or information related specifically to ZIP codes as they existed in the two earlier decades we refer you to the original pair of documents.
The New 2010 ZCTAs
We first we saw of the 2010 edition of the ZIP Code Tract Areas was in the spring of 2011 with the release of the SF1 (Summary File 1) data product. As they did in 2000, the Bureau defined the new ZCTA's as an aggregation of the current census blocks. Which are now, of course, 2010 blocks. They updated their ZCTA geography page to reflect the new version. They noted that the experiment with creating the HH and XX pseudo-ZCTAs was dropped for this decade. Probably because they caused so much confusion among users. Readers of this page (previous link) should focus on the paragraph beginning It is important to note the following: .
New (Post-2010) ZCTA Master Data SetAn important resource created by the MCDC is the master data set where we combine geographic and demographic data regarding ZCTAs. The version of this data set created to describe the 2000-vintage ZCTAs has been renamed to zcta_master07. We have replaced (i.e. written over with new data) the zcta_master data set so that it now describes the 2010 ZCTAs, and contains demographic indicators taken from the most recently available American Community Survey data. We initially created this new version in 2012, and it used our estimated ZCTA level data allocated from 2006-2010 tract-level period estimates. In August, 2013 we did a major update of the zcta_master and used official 2007-2011 ZCTA-level ACS data for the first time. The data set now contains the current Congressional District codes (113th congress) as well as the updated ("2012") PUMA codes. We also have updated the various metropolitan area codes: cbsa (Core Based Statistical Area, i.e. Metropolitan Statistical Areas and Micropolitan S.A.s), Metropolitan Divisions and Combined Statistical Areas (CSAs).
New United States Zip Codes Web SiteWe were recently referred to a new web site - http://www.unitedstateszipcodes.org/ which appears to overlap extensively (in terms of content) with our own "All about ZIP code..." pages. But it includes some ZIP lookup and mapping tools that ZIP code users should find very useful. It also features information about shipping rates between ZIP codes for USPS, FedEx and UPS. It has a free downloadable database, but it also has one that it sells for $29.95. We have not yet had time to use this site much but it certainly has the look of a very useful resource.
2010 Decennial Census Data
Summary File 1 contains data tables summarizing the results of the 2010 decennial census survey. (The 2010 census did not include a long form, only a short form.) Data for ZIP codes (or ZCTA's, to be move precides) appear with Summary Level codes of 871 (ZCTA within state) on the individual state files. The national file has summaries for all ZCTA's in the nation and has three summary levels: 860 (complete ZCTA's); 870 (ZCTA's within state - same data as the 871 summaries on the state files but with a different SumLev code just to confuse you); and 880 (ZCTA within county). Since over 90% of all ZCTA's are contained within a single county you mostly get three copies of the same tabular data on the national file. These data can be accessed via the American FactFinder query application at the Census Bureau, or they can be accessed from the Missouri Census Data Center's public data archive using the Uexplore/Dexter query tools. Specifically, go to (sf12010 to access the full tables or sf12010x to access the much smaller (but often more useful) standard extract data. The latter can also be accessed using our Standard Summary File 1 (2010 Census) Extract Assistant web app. The SF1 Census Profile lets you generate formatted profile reports for up to 4 geographic areas at once, with ZCTA among the levels available.
American Community Survey Data
If you are not familiar with the ACS we recommend you see our American Community Survey page for basic background information. Basically, ACS does for 2010 data what SF3 did for 2000 data. Except that it is not simple one-point-in-time data derived from a census, but instead uses survey data collected over a 5-year period (2008-2012 is/was the latest available as of Jan, 2014). New data are normally released in December each year.
You can access these data from the American FactFinder web application at the Census Bureau. You can also access ZCTA data from the MCDC's data archive using Uexplore/Dexter, or by viewing data for ZCTAs in profile report format. Here are the key links:
Just as we have 3 versions of ZIP Code Resources documents for three decades we now have three versions of the MABLE/Geocorr geographic correspondence engine for the same three decades. The new version is at /websas/geocorr14.html. You can use this to relate the new codes to all the others such as counties, metro areas, urban/rural, PUMAs (both old and new), 2010 census tracts, congressional districts, etc. (This version of MABLE/Geocorr replaces the interim 2010 version which is no longer supported).
Correlation List Files Relate ZIP Codes to Common GeographiesThe previous paragraph describes the MABLE/Geocorr tool which you can use to generate "correlation lists" - files that define the correlation of one type of geography (such as ZCTAs) to another. We have used this tool to create a directory of commonly requested equivalencies where the "Source" gecode is the 2010 ZCTA and the target geographies vary for each file in the directory. You can access this collection at /cgi-bin/uexplore?/pub/data/corrlst/zip2_xxx .
Initially we have stored files relating 2010 ZCTAs to
Resource Relates ZIP Codes to ZCTAs
NOTE: This resource is based on the Excel spreadsheet posted by John Snow, Inc., which we downloaded and stored as a .csv file (referenced below) in the fall of 2015 (updated 11-2-2015 from previous version that was vintage 2010). You can "go to the source" for these data at http://udsmapper.org/zcta-crosswalk.cfm. (They may be later than 2015 when you access in 2016 or later.)
We sometimes like to think of ZIP codes and ZCTA's as more or less the same thing, the only difference being that ZCTA's are "rounded off" to census blocks while real ZIP codes are not. That is more or less true when talking about standard residential ZIP codes. But there are others; ZIP codes that are neither "standard" nor residential. The ZIP code for Reader"s Digest, for the North Pole (Santa Claus), for the University of Missouri, for Camp Pendleton, for Google headquarters, etc. These are the special (aka "point", "unique", etc.) ZIP codes that are important to the post office and to the entities that use them to receive their mail. But they are not spatial areas (they are points in most cases) and people do not live in them. The ZIP code 65211, for example, is assigned to the University of Missouri in Columbia, MO. If you want to send me a package here at my office it will be expedited if you use the 65211 ZIP code. But you can also use ZIP 65201 and the folks at the post office will know to put it in the 65211 cubbyhole because they just know where it is. If there were no 65211 then 65201 is the residential ZIP code where the campus is located so that it is what would be used.
A new dataset stored in our public archive provides a crosswalk between all (or at least mostly all) current ZIP codes and the ZCTA's to which they correspond. For standard residential ZIP codes it's simple; they are the same codes. For ZIP code 65211 the ZCTA is 65201. You can either access the original downloaded csv file or you can use Dexter to access the converted dataset (zip_2_zcta15) . We created an alternate version of these data in the form of a SAS format module (that lets us do easy table lookups within SAS data steps, which is quite handy if you are a SAS programmer, but totally not relevant if you are not.)
HUD and USPS Provide Links Between ZIP Codes and Census Tracts
The HUD-USPS ZIP Crosswalk files provide a very useful tool for relating ZIP codes and census tracts. These are actual true ZIP codes (not ZCTAs, and including all the "special" ZIPs such as point ZIPS and P.O. box ZIPs). These files are available quarterly going back to 2010. Beginning with the first quarter of 2012 the tracts are 2010 versions; before that the tracts were vintage 2000. The web site (linked to just above) allows you to choose a quarter and a "Crosswalk Type" (ZIP to tract, or tract to ZIP) and download an Excel file with data for the entire country (including Puerto Rico). Each record/row of the file specifies a 5-digit ZIP code and a census tract (including State and County codes). The degree of intersection between the 2 geographic entities is measured using a set of 4 ratio variables. RES_RATIO specifies the portion of all residential deliveries within the ZIP code which are in the tract (this is on the ZIP-to-tract type file; on the Tract-to-ZIP type this variable measures the portion of the deliveries within the tract that are within the ZIP). BUS_RATIO is the same concept but only works with business addresses. OTH_RATIO handles addresses that are neither residential nor business. Finally, TOT_RATIO measures the portion of all addresses regardless of type. That's it; very simple. Note that these are all ratios with no actual counts. That would be nice, but then the USPS would have a hard time selling its ZIP Delivery Statistics data products.
The intened use of these crosswalks are for users wanting to allocate data between the two types of geography. Use the ZIP-to-tract correspondences/ratios to allocate data already available at the ZIP level to census tracts. Use the tract-to-ZIP to allocate data at the tract level to ZIP.
The Missouri Census Data Center has downloaded the data for the 1st quarter of 2016 (more later, if user interest dictates) and made them available in our public data archive. These can be accessed using our Uexplore/Dexter software in the new hudusps data directory. The two crosswalk files for 1st quarter 2016 are in the ziptract2016q1.sas7bdat dataset. We merged the two crosswalk files (the original files in csv format rather than xlsx are available in the directory as well) have been merged to create the single data set. We just have two sets of ratio variables, one for going ZIP to tract and a second set going the other way. E.g. variable bus_ratio2 is "Portion of tract business addresses in ZIP >" while bus_ratio is "Portion of ZIP business addresses in tract".
We wanted to try out this new toy so we looked around to see what data we had at the ZIP level. We have lots of census data at the ZCTA level (almost all of which is also available for tracts) but not much by true ZIP codes. We do have one: the ZIP Business Patterns data published annually by the Bureau. So we went and did the minimal programming needed to apply the apportioning ratios to the ZBP data to create at new "tract business patterns" data set for the entire U.S. See the results in the ustractsbpests.sas7bdat dataset. Follow the link to the metadata for this set. If you have any interest in seeing the SAS(r) code we used to create it; look for the line The dataset was created 13JUL16 by the allocate_zbp_totract SAS program stored in /pub/data/hudusps/Tools and follow the link associated with the program name.