13 replies to this topic

[benbakelaar]

"I believe this was my source: [url="http://"http://civicspacelabs.org/home/zipcodedb"]http://civicspacelab.../home/zipcodedb[/url]
"

[gregsd]

"Here's another free source to check out [url="http://"http://www.thedigitalmapstore.com/Free_Stuff/ZIP_Code_Database/zip_code_database.html"]http://www.thedigita...e_database.html[/url], if you haven't already.

Greg "

[Nick Springer]

There are a few inherent problems with Zip Code mapping.

For
one, Zip Codes are technically not polygons. They are sets of addresses
and apparently there can even be a mixture of Zip Codes along the same
street in rare cases.

Two, some Zip Codes are point features (many large building have their own Zip Codes).

Three,
the USPS changes their Zip Codes basically all the time, and they don't
publish the information in a centralized way. Zip Codes are added,
removed, and redefined all the time. So, even data from 2000 is going
to be pretty outdated.

Unfortunately I don't have a good solution for you. I recently did a Zip Code based territory map and had issues.

[Martin Gamache]

Nick is correct.

Zip codes get updated officially every 30 to 60 days by the USPS.

[benbakelaar]

Thanks
for the explanation Nick. So even big companies that run zip code
searches, basically don't have anything more exact than what is
publically available?

[Kartograph]

No,
they can pay hard cash for it. The USPS offers a subscription service:
You pay loads of money and you get all the updates for your database
once they become available.
BTW, ZIPcodes can even consist of two disjunct areas in very rare cases.

[benbakelaar]

"Thanks
for that info as well. I've seen the disjunct areas myself actually, if
you all haven't seen this tool yet, it's worth checking out.

[url="http://"http://maps.huge.info/zip.htm"]http://maps.huge.info/zip.htm[/url] "

[benbakelaar]

"Maybe one of you can explain this:

""""
Can I get a file that cross-reference ZIP Codes with latitudes and longitudes?

The Postal Service does not maintain ZIP Code maps for sale to the public.

The
TIGER/ZIP+4 File relates ZIP+4 codes to Census Bureau data (e.g.,
latitude, longitude, Standard Metropolitan Statistical Area (SMSA),
etc.). For more information about the TIGER/ZIP+4 File, call the
National Customer Support Center at (800) 238-3150, select Option 6 and
then Option 2.

In addition, there are many commercial software
products available with geographic information. Check with your local
software store or search the Internet for related Web sites.
Back to Top
""""

On https://hdusps.esecu...view=1#latitude

Linked from [url="http://"http://www.usps.com/ncsc/ziplookup/lookupmenu.htm"]http://www.usps.com/.../lookupmenu.htm[/url] "

[Nick Springer]

"

Thanks 
for that info as well. I've seen the disjunct areas myself actually, if 
you all haven't seen this tool yet, it's worth checking out.

http://maps.huge.info/zip.htm

http://

...and to prove my point, the Zip Code polygon they show for 08515 (my zip) is at least 5 years out of date. "

[benbakelaar]

"This
is a great learning experience for me, I knew geographic data was
always subject to change, but this really highlights the issue well!

I've done my research, and here is what I've found:
The
United States Postal Service offers Address Information System (AIS)
Products, only one of which contains latitude and longitude
coordinates. That is the TIGER/ZIP+4, product ID ZP260R. I spoke with
someone at National Customer Support Center (NCSC) 800-238-3150 and
they explained that this product is based on Census data, which is
updated less than yearly (they weren't really sure about the exact
frequency) and only shows ""matched"" data, which means out of a total of
40000+ zip codes, this TIGER/ZIP+4 file may contain about 27000
""matched"" zip codes with lat/long coords. The product costs $700 for
all states, one-time purchase, or $50 per state.

Now I'm wondering how accurate the Address Locator I've created in ESRI ArcMap 9.1 is! "

[Kartograph]

"AFAIK:
The
census ""ZIP"" code areas were created as non-overlapping polygons GUIDED
by the ZIP codes of 2000. So they are handy for lat/lon and ZIP-code
geographic analysis, but I would surely not send letters with this
data. "

[benbakelaar]

"

AFAIK:
The 
census ""ZIP"" code areas were created as non-overlapping polygons GUIDED 
by the ZIP codes of 2000. So they are handy for lat/lon and ZIP-code 
geographic analysis, but I would surely not send letters with this data.

http://

Actually,
I guess I didn't mention it before, but that's what the outcome of one
of my projects is. We are sending a targeted direct mailing to retired
physicians within a certain radius of various free clinics, to
encourage them to volunteer. So, absolute exactness is not essential,
but it would be nice to have some confidence that the radius search was
reasonably accurate. Any comments on that? "

[Kartograph]

I
think if car mobility is concerned and you just want to check if a
given ZIP are isn´t too far away, go with the Census Data. The Errors
should stay single digit in miles. But someone else might be better
informed.

[loximuthal]

"For
complete answers to what Census Bureau ZCTAs (ZIP Code Tabulation
Areas) are and what issues they have, you can take a look at the [url="http://"http://www.census.gov/geo/ZCTA/zctafaq.html"]ZIP Code Tabulation Area FAQs[/url].
While I have not been involved with these directly, I do hear a bit
about them at various meetings here at the Census Bureau. Any location
errors should be limited to the size of a single block. I offer that
with the caveat that some rural blocks are rather large (some in
Western mountainous areas are larger than many Eastern counties), but
for the most part the vernacular understanding of ""block"" corresponds
well to the Census block. "