codeforamerica / address-normalizer

A simple tool that takes a CSV file with addresses and normalizes them.
BSD 3-Clause "New" or "Revised" License
16 stars 4 forks source link

What does a normalized address look like? #3

Open mick opened 11 years ago

mick commented 11 years ago

Should we define our standard of normalized address? (or is there a good example somewhere already?)

open questions: 2 or 3 letter address type abbreviations (ave, st) proper street name capitalization (not important for matching, but important for display) *zipcodes +4 digits?

other questions?

atogle commented 11 years ago

I'm not sure about the context here, but the USPS has a web service for address normalization that could be useful, either for defining standard addresses or just to use. Not sure about the terms/limits.

https://www.usps.com/business/web-tools-apis/address-information.htm

mick commented 11 years ago

The plan is to make use of either Easypost or USPS APIs for a subset of the data (a unique set of street names / types) within large datasets, and use those responses to normalize the rest of the dataset without query the API every record.

I agree we would normalize to what USPS uses.

daguar commented 11 years ago

Agree with both of you in using USPS-based -- my understanding is that the EasyPost API most likely wraps around the USPS one (perhaps with a caching layer), but maybe we test that assumption?

@atogle The rationale for using EasyPost right now is that, ostensibly, usage is free/without limits that would affect our purposes: http://blog.geteasypost.com/post/40684464899/freeaddressverification

bensheldon commented 11 years ago

You might also check one of the commercial "change of address"-companies to make sure that your format works with them too (I assume it should because USPS, but might be good to check). 

If you don't know what I'm talking about, these are companies where you upload a list of addresses (like your nonprofit's list of donors, or, ahem, direct mail marketing) and they  will update the addresses if those people have submitted a Change of Address to the USPS.

On Fri, Apr 12, 2013 at 8:39 AM, daguar notifications@github.com wrote:

Agree with both of you in using USPS-based -- my understanding is that the EasyPost API most likely wraps around the USPS one (perhaps with a caching layer), but maybe we test that assumption?

@atogle The rationale for using EasyPost right now is that, ostensibly, usage is free/without limits that would affect our purposes: http://blog.geteasypost.com/post/40684464899/freeaddressverification

Reply to this email directly or view it on GitHub: https://github.com/codeforamerica/address-normalizer/issues/3#issuecomment-16300056