I have Postcode in my large database, which contains values like SL5 9JH, LU1 3TQ etc.
Now when I am pasting above postcode to maps.google.com it\'s
A quick note for those finding this SO answer. The answer by Daniel Vassallo uses the Google Geocoding API V2 which has now been deprecated. The new v3 API uses a request format like this:
http://maps.googleapis.com/maps/api/geocode/output?parameters
An example for a postcode lookup, returning the data in JSON format is:
http://maps.googleapis.com/maps/api/geocode/json?address=SL59JH,+UK&sensor=false
This returns a JSON array that includes the lat and long in results->geometry->location->lat and results->geometry->location->lng
Example response:
{
"results" : [
{
"address_components" : [
{
"long_name" : "SL5 9JH",
"short_name" : "SL5 9JH",
"types" : [ "postal_code" ]
},
{
"long_name" : "Windsor and Maidenhead",
"short_name" : "Windsor and Maidenhead",
"types" : [ "administrative_area_level_2", "political" ]
},
{
"long_name" : "United Kingdom",
"short_name" : "GB",
"types" : [ "country", "political" ]
},
{
"long_name" : "Ascot",
"short_name" : "Ascot",
"types" : [ "postal_town" ]
}
],
"formatted_address" : "Ascot, Windsor and Maidenhead SL5 9JH, UK",
"geometry" : {
"bounds" : {
"northeast" : {
"lat" : 51.39655490000001,
"lng" : -0.66024660
},
"southwest" : {
"lat" : 51.39457330,
"lng" : -0.6624574999999999
}
},
"location" : {
"lat" : 51.39539040,
"lng" : -0.66096740
},
"location_type" : "APPROXIMATE",
"viewport" : {
"northeast" : {
"lat" : 51.39691308029150,
"lng" : -0.6600030697084980
},
"southwest" : {
"lat" : 51.39421511970851,
"lng" : -0.6627010302915021
}
}
},
"types" : [ "postal_code" ]
}
],
"status" : "OK"
}
The API spec is available here: https://developers.google.com/maps/documentation/geocoding/
The technical term for the process you describe is called reverse geocoding. Google offers the The Google Geocoding Web Service New working Google Geocoding Link, where you can do reverse geocoding on the server side, instead of in JavaScript on the client-side.
For example, if you try the following URLs in your browser, you would get back the latitude and longitude of the postcode passed in the q
parameter, in CSV format:
http://maps.google.com/maps/geo?q=SL59JH,+UK&output=csv&sensor=false
http://maps.google.com/maps/geo?q=LU13TQ,+UK&output=csv&sensor=false
This is how you would be able to reverse geocode your postcodes in php, for example:
$url = 'http://maps.google.com/maps/geo?q=SL59JH,+UK&output=csv&sensor=false';
$data = @file_get_contents($url);
$result = explode(",", $data);
echo $result[0]; // status code
echo $result[1]; // accuracy
echo $result[2]; // latitude
echo $result[3]; // longitude
Note that as Pekka suggested in another answer, the Google Maps API Terms of Use seem to prohibit the storage of the results, unless the store acts as a cache for data that will used in Google Maps. You may want to get in touch with Google and enquire on the Google Maps API Premier to have more flexible terms of use for your geocoding requirements.
The Ordnance Survey have released the postcode locations on a Creative Commons licence (CC-BY v3, IIRC). It would be a lot less hassle (and a lot clearer legally) to use that instead.
There's even a version with WGS84 (a.k.a. GPS) coordinates mirrored by mySociety
Check out ALgolia places here. Super fast, Open source and MIT
The Google Geocoding API does that, although if I remember correctly, their terms of service forbid local storage of the geocoding results.
I know this is an old question, but just chipping in here with how I managed to achieve the same thing (in PHP, but should be fairly simple):
I had a database of thousands of differently formatted postcodes. I cleaned each and every one of them up, uniformly, with this function and batch updates:
function clean_postcode($postcode)
{
$out = preg_replace("/[^a-zA-Z0-9]/", '',strtoupper($postcode));
if(strlen($out)>3)
{
$out = substr($out, 0, (strlen($out) -3)).' '.substr($out, -3);
}
return $out;
}
Now that all postcodes are formatted uniformly, I downloaded and imported the Ordnance Survey's Code-Point data set (free) - https://www.ordnancesurvey.co.uk/opendatadownload/products.html
I imported all their CSVs into my database in a separate codepoint
table. From each CSV I imported the postcode, and the Eastings and Northings values.
I ran the clean_postcode()
function again in batch on all the Ordnance Survey data in my DB. About 50% of the postcodes have spaces in, and 50% don't - after this they were uniformly set.
I ran the following PHP script on each and every postcode in the codepoint
table and saved the Latitude and Longitude values returned into this table: http://bramp.net/blog/os-easting-northing-to-lat-long
All done! You can now match up and pull a Lat/Lon value based on well-formatted postcodes.
Further reading: http://www.deepbluesky.com/blog/-/converting-os-coodinates-into-longitude-latitude_7/