|
I need to test a search index for international addresses. I need a sample dataset of 100 addresses in local language for every country (or as many as possible). I don't need it broken down to address components or geocoded. Just a random set of street addresses for the world in what I call 'full address' format: 1 microsoft way seattle wa xxxx are three example address (The boshoi theater is the last one) |
|
The addresses are inside the planet.osm dump ( http://planet.openstreetmap.org/ ), but you'll have to extract/parse them yourself. Address tag format is described at http://wiki.openstreetmap.org/wiki/Key:addr . Probably you could also get something out of overpass API ( http://wiki.openstreetmap.org/wiki/Overpass_API ) , but I am not familiar with that. This is very useful information. I appreciate your answer to my question a LOT! When I get the data extracted, I will put it up on github
(03 Mar '13, 15:44)
gearond
I found that the only real data in the OSM database is the geo coordinates. I will actually have to query google for reverse geocoding and process that. Looks like I have some script writing to do.
(03 Mar '13, 17:37)
gearond
I will have to get the geocoordinates from osm, then reverse geocode them with google (after setting my browser to the local language) to get the formatted address string (full address) that I'm looking for.
(03 Mar '13, 17:42)
gearond
I have some code being written that will get me this database. I will publish it to github when it's done.
(10 Mar '13, 15:01)
gearond
AF;ps_AF;سرک کولو له پشته، کابل، افغانستان;34.537986;69.164316;place_of_worship; AF;ps_AF;جاده صلح، کابل، افغانستان;34.517789;69.189979;place_of_worship; AF;ps_AF;جاده نادر پشتون، کابل، افغانستان;34.515477;69.180446;place_of_worship; AF;ps_AF;سرک کولو له پشته، کابل، افغانستان;34.539130;69.165019;telephone; AF;ps_AF;سرک کاریز، کابل، افغانستان;34.506858;69.123537;place_of_worship; AF;ps_AF;Seh Aqrab Road, کابل، افغانستان;34.513626;69.147141;place_of_worship; AF;ps_AF;سرک چهارم انصاری، کابل، افغانستان;34.532946;69.160152;hospital;
(16 Mar '13, 02:18)
gearond
Above is sample output, of what will be in the list of geocoded addresses in the local language. Right to left text really messes up text editors.
(16 Mar '13, 02:19)
gearond
showing 5 of 6
show 1 more comments
|
|
An alternative about using the overpass-API is to download some OSM data extracts for countries or part of each that you can find on geofabrik.de or other sources according to /wiki/Planet.osm Then do a filtering with the tool called osmfilter about tags beginning with "addr:" If you need your data set in text form, try osmconvert and its CSV conversion feature. |
|
Non OSM related answer, but the Royal mail website has a short list of correctly formatted international addresses http://www.royalmail.com/personal/help-and-support/Addressing-your-items-Rest-of-the-World and here http://www.royalmail.com/business/help-and-support/addressing-your-items-western-europe Please note that for Belgium, the house number does not always consists of only numbers: 12, 12a, 12b1, 12/1, 12 bus 4 are all perfectly possible house numbers
(01 Mar '13, 04:42)
escada
The issue that I have with those answers is that they are all in English. I want to serve all people with their langauge the way that google would serve them addresses.
(02 Mar '13, 13:54)
gearond
|
I'll look at the un commenteted answers tonight, thank you all.