ZIPs from Scan/US
The Scan/US ZIPs, consisting of both polygon and centroid representation (centroids show enclosing ZIP as well), are the most accurate that you can obtain at any price. And they are reasonably priced, at $995 for single user (please call 800-272-2687 for commercial license pricing).
Up-to-date at the current year, these ZIP boundaries match the full range of demographic data, and population characteristics, that are available from the Scan/US data catalog. (The demographic data is also available for purchase — that's the "data" part of mapdata2go.com!)
The ZIP code, inaugurated into service on July 1, 1963 by Postmaster General J. Edward Day, is a territory designed to help in delivering the mail.
ZCTA (Census 'ZIP' codes)
The Census Bureau made pseudo ZIP codes called ZCTAs. Meant to provide geographic areas corresponding to mailing areas for households, the 2010 Census ZCTAs became available in 2013. To quote the Census bureau on this, "ZIP Codes assigned to businesses only or single delivery point addresses will not necessarily appear as ZCTAs," and "Some addresses will end up with a ZCTA code different from their ZIP Code."
But ... ZCTAs are free with no development cost, and for this reason many companies like to give them away.
Since we make our own ZIP boundaries, we know we can make them better, as well as more up-to-date, than the ZCTAs.
How we make great ZIPs
In the next panel, you can read about how we make ZIP boundaries and why we believe our ZIP demographic data is more accurate than our competitors' data
Scan/US ZIP Code Methodology
ScanUS develops ZIP code cartography in-house, for semi-annual release. The ZIP code layer is a dynamic cartography: it changes according to the USPS's decisions on mail distribution logistics.
Thus, the source material for the ZIP boundaries (and ZIP point locations) is the most recent TIGER geographic files and detailed TIGER streets information, combined with current USPS address range data and supplemental ZIP+4 coordinates for new residential developments. (Scan/US uses all of these information sources to produce other data products as well, including the Scan/US Geocoder, postal carrier route locations and ZIP+4 locations)
Combining this raw data, ScanUS links ZIP code information to sub census block cartography. In other words, map components more detailed than census blocks are linked to ZIPs. Then, we apply techniques of computational cartography such as setback corridors, medial axes and voronoi diagramming to these low-level mapping units. They are carved up and merged together to yield ZIP code boundaries that match how mail is delivered at that time.
A benefit of this technique [in addition to unparalleled accuracy – and timeliness – of the ZIP boundaries themselves] is a superb delivery-count-weighted crossreference between ZIP code areas and census areas. By knowing which streets in a block belong with which ZIP+4 and how the data in a block is distributed among the street based on delivery counts for each ZIP+4, Scan/US knows the relationships between ZIPs and all levels of census cartography.
This allows us to bring current year demographic data to the ZIP code level easily and accurately. The same is true of 5-year projection data. Perhaps most surprising, though, is the way we use it to assign year-2000 (or 1990) counts from the decennial US Census to current time-period ZIP codes. This little bit of cross-reference magic makes it easy to compare historical trend data on a geographic area of current commercial value (today's ZIP code) that would not otherwise be possible. Year-2000 and Year-2010 population characteristics in current-year ZIP codes. Scan/US does this, and we can totally back it up.
Scan/US does not need to rely on such inaccurate methods as block or blockgroup centroid matching nor on distribution by fractional area overlap. Either of those techniques, no matter how competently applied, unavoidably leads to dramatic inaccuracies accumulating data at the boundaries of the ZIP codes, which is why we do not use them.