Where can I get sample data with valid addresses?

December 7, 2007 8:21 AM

Where can I get sample data with valid addresses?

A friend of mine is testing an application that incorporates Google Maps API, which doesn't seem to map their location without a valid exact address. He's looking for some test data to use with 1000 records, 20 valid addresses from each state. We've tried the Fake Name Generator, but that didn't seem to generate valid addresses.

Can anyone recommend a good utility to use for this?

posted by aranyx to computers & internet (4 comments total)

There are lots of direct marketer resources who, I presume will sell you real addresses.

You could find a city with rational layout, pick a really long, straight street and get the address of the first house on the block, then write a script to add one two thousand times. That will probably give you a decent amount of valid addresses.

It is good to test the invalid case as well.
posted by jeffamaphone at 8:44 AM on December 7, 2007

There's a CSV list here with the latitude/longitude and street address of all Starbucks locations (the "Starbucks No Linefeeds.csv" inside the archive). I'd be pretty surprised if that didn't have 20 valid addresses per state. You'll have to parse the address out, but it'd be a pretty simple job.
posted by uncleozzy at 8:58 AM on December 7, 2007 [1 favorite has favorites]

Uncleozzy, that should do exactly what we're looking for, thanks!
posted by aranyx at 9:02 AM on December 7, 2007

Not sure if this application is already using google maps api for the geocoding, but if not, you may find that it gives you more valid co-ords. IME, gmaps will pretty much give you a valid location (or near valid) with as little as a state and zip -- in most situations, it will automatically 'genericize' the address in order to give a valid co-ordinate back, and if it fails, you can always strip the street and chuck it back in.

there is a latency cost of going to google to get geocodes, but depending on how many addresses you're mapping, it's not terrible, plus it's pretty trivial to cache them.
posted by fishfucker at 9:21 AM on December 7, 2007