[wplug] more about zipcodes

[Eric Cooper]

Peter Lucas of MAYA is interested in and knowledgeable about this kind
of thing, so I forwarded the original messages to him.  His response
might be of interest to the list:

----- Forwarded message from Peter Lucas <lucas at maya.com> -----

Date: Sat, 02 Jul 2005 13:11:11 -0400
From: Peter Lucas <lucas at maya.com>
To: Eric Cooper <ecc at cmu.edu>
Subject: Re: [wplug] OT: Download data with zip code and long/lat?

Eric,

As the text at the referenced Census links correctly imply, zip codes
are a lot more complex (both technically and politically) than they
appear to be at first glance. Examples:

1) They don't really refer to geographic areas. They are merely
routing codes, and are often topologically complex.

2) They change constantly to accommodate the needs of the post office
to get the mail delivered. The Post Office is sensitive about
maintaining this flexibility and doesn't want to get trapped into
being forced to make stability guarantees (the way the DOD did with
GPS, for example).

3) One has to worry about 5 and 9 digit zip code issues. Some
low-population-density areas actually have *3* digit zip codes!

4) The Post Office (which, remember, is quasi-private and
self-supporting) considers zip-code data a profit center and didn't
much like it when the Census Bureau gave it away.  A compromise was
reached via the introduction of the ZCTA concept:
    http://www.census.gov/geo/ZCTA/zcta.html
which is a kind of a thin abstraction above zip codes tuned for
demographic purposes.

The bottom line is that free, up-to-date zipcode geographic data is
not currently available.  There are many companies that license these
data from the Post Office and repackage it for sale. Example:
    http://www.zip-codes.com/zip_database.asp

But you won't find it for free.

Very precise (but not  completely up-to-date) 9-digit zip code data is
implicit in the raw Tiger dataset, but would require a lot of
processing to extract.  Civium will probably do this work eventually
and it will be made freely available, but it is not particularly high
priority. This is exactly the kind of thing that Civium is for--to
liberate data that is legally public domain but effectively
unavailable for non-constructive reasons.

As for the larger question of sources for geographic data, that would
be a much longer email message. There is no shortage of data, the
challenges are all in the area universal identity and data fusion. As
you know, that is exactly what we are working on in a big way.  If
Branen wants to hear more, he'd be welcome to come and visit.

Feel free to pass this along, if useful.

----- End forwarded message -----

-- 
Eric Cooper             e c c @ c m u . e d u