ZIP Codes are Not Areas

ZIP codes are postal codes in the United States created by the US Postal Service. Although the Postal Service refers to them using an all-caps word, virtually everyone else calls them Zip codes.   Zip codes also appear as longer codes called Zip + 4 codes.   In this note, we will refer to both Zip codes and the longer Zip+4 form as Zip codes.


A very common misconception in GIS is that Zip codes identify  polygonal regions or areas. People often think of mapping in the US as a hierarchy of ever-subdivided polygonal areas: states, counties, cities, zip codes. If they need higher resolution than a county, they next leap to zip codes because they think of zip codes as polygonal areas.   But that is a mistake, since Zip codes are not areas.  


Zip codes usually are linear features associated with specific roads or are point features associated with specific addresses such as apartment buildings or military bases.   In some cases, Zip codes have no physical location because they are assigned to a mobile or abstract "location" such as a military ship, and thus they are not even points.


Even in the most common case of Zip codes assigned to streets, Zip codes do not clump always together in groups that may be covered by rational polygons. Frequently, linear features to which the same Zip code are assigned will be outliers that are not located in the same contiguous clump of linear features assigned that same Zip code.


We can consider an example using a map of part of Reno, Nevada, shown below.



This map is fairly typical of the situation in mid-sized urban areas. It is extracted from the US Census Bureau's TIGER/Line 1997 data set, which includes roads as segments of lines, with most line segments coded with Zip codes for that particular segment.


To create areas from road lines where lines have a common zip code there are several approaches. One possible approach is to select all line segments with the same Zip code and to then draw an area object that encloses them. This can be done by creating a buffer zone about each street line having a particular Zip code and then doing a Union of the buffer zone areas thus created. The blue, purple, and green areas were created in this way and each represent a a different Zip code value.


The road lines shown in red selection color all have yet another Zip code in common, which is different from the Zip codes used to create the blue, purple, and green areas. Immediately there are three pathologies visible in this map:





The above situation is very common.  Almost any urban map in the US will show similar, if not even more bizarre effects. Rural maps can have such a sparse network of roads with such strange zip code assignments that some rural areas cannot even be approximated with zip code regions.


For the above reasons, any map that purports to show "Zip Code Areas" or "Zip Code Polygons" should not be taken as a precise map showing Zip code locations. It is at best some sort of approximation and most likely is wildly inaccurate in certain regions. The approximations might be useful, but they should not be confused with the real thing.


The US Postal Service, of course, does not make it any easier to deal with such issues by making it easy to get Zip code information. Zip code information is not available for download via Internet from the US Postal Service. It is best obtained from the US Bureau of the Census.


ZIP Code Tabulation Areas (ZCTAs)

For statistical tabulation purposes the Census Bureau has long found it convenient to work with Zip code groupings of population. Zip codes have been so useful that the Bureau embarked on a project to create a standardized map of the US showing the approximate region of coverage of various Zip codes as areas. These areas are known as ZIP Code Tabulation Areas (ZCTAs). ZCTAs may be downloaded from the Census Bureau's site. Drill down to the Cartographic Boundaries pages to get the ZCTA data set provided as part of the Census Bureau's TigerLine series. Download them using .e00 format so they will import into Manifold using the correct NAD83 datum.


Before ZCTAs were published, every vendor of maps used in GIS had to resolve the various ambiguities posed by Zip code pathologies like those shown above. With ZCTAs the GIS industry now uses a standard approximation that is the same used as the Census Bureau for publishing demographic information.




The problem with ZCTAs is that they do not cover much of the country.  In regions where there are few zip codes, ZCTAs do not exist, as seen in the image above in much of Nevada and in mountainous regions of California.


See Also