I currently have a mySQL table of addresses of certain locations that I need to obtain Latitude and Longitude for. This is a table being used in WordPress to map store locations to google maps. I am looking for the best way to obtain the lat/lon for all of the records in this database.
Does anyone know a simple strategy I could use to run through all of the records and obtain the lat/lon of each address?
Google have the Geocoding API - http://code.google.com/apis/maps/documentation/geocoding/
Not sure what kind of addresses you mean. If IP then there are databases that will give you the approx lat and long for most of your visitor's IPs. You will have to capture it then.
If you mean postal addresses then, I know of nothing to help world wide. In the UK, the Royal Mail's PAF used to contain lat and long, but you must have the post code. I suspect there might be similar data available for ZIP codes in the US and Cedex codes in France.
Other than that, I expect you will be out of luck.
Are you asking for code? I use a bash script to solve it. It's not a copy actually it is quite simple with Linux bash commandl-ine. Most likely you want to dump your database into a text file and then read and query it from them shell. It took me about 4-5 hours for a country like Germany with postcode from 0-99999.
Related
I would like some help with how to go about calculating the distance between two post codes. Here is the setup that I wish to have, let me give you a scenario.
A user logs into my website (they have provided their details, their postcode being one of them). That data is stored in the users table.
I have another table with details about locations. I would like to display locations (and the distance) that are closest to the user.
I know that there are certain methods that can be used, for example:
Converting a postcode to a long-lat and using built in methods to calculate the distance, I think Google offer a free service. But I have no idea how I would implement such a thing. All the tutorials I have taken a look at seem too complicated and for the time being I want a very crude and simple solution upon which I can build and more importantly learn!
I am fairly familiar with PHP and MySQL when it comes to basic input/output of data. But I am not sure of the best way to handle the information with regards to user data, and the locations of the 'businesses'.
Any suggestions, comments will be more than appreciated.
If you're US based, this API should meet your needs: https://www.zipcodeapi.com
It's made specifically to find distance between two zip codes, tells you the long-lat for zip codes, find all zips relative to a location, etc.
Get postcode northings eastings data free from ordinance survey (credit them on your webpage to get free)
2 calculate distance from the appropriate csv flat file (M.csv contains all manchester postcodes and north and eastness).
Distance is sqrt((n1-n2)**2 +(e1-e2)**2) - excuse my fortran
If its slow,and you want an approximate distance create a small database of first half postcodes averages so you are looking up M20 not M20 3nb.
5 NB google don't like you using their postcode apis outside google Maps.
I am wondering about an algorithm so that for short distances, if you are looking at places close to a postcode, you just search M postcodes for a M postcode (and maybe nearby post areas like SK, OL, if you have a table of adjacent postcode areas).
7 Any body taken this further?
I have a bunch of messages (from twitter) that include addresses. They are in various kinds (as many as you could imagine a random sampling of people enter an address. The city location is always known so they normally just put a road name and number/area)
Is there any library out there to extract these? I've tried looking but found nothing.
If no, any suggestions as to how I do this? At the moment I am just extracting things like [previous word + [rd/ave/street/lane/blvd]] but it isn't that accurate.
Any ideas?
Thanks
I know of no library that does this.. but a crazy idea came to mind while reading your question.
Use the google maps geocoding api to find long and lat for your address..
then use the reverse geocoding api to find the address from your lat and long since it will be neatly formatted in a json object.
Quite messy but it is the best i can come up with. (Has the upside that you then already have the coordinates of your address :)
I am working on a project that I have array of GPS locations.
I want to know for each of those points, where are they?
I have a specific list that I am trying to match their location to one of those such as: Water, Highway, railway, ...
I tried this API (http://developers.cloudmade.com/wiki/geocoding-http-api/Documentation) but it doesn't give me any accurate data! even less than 50% correct!
Do you know any better solution?
I use php for server side that is responsible to gathering data from different webservices.
Our client sides are Android, Java
I expect system search and find the type of location automaticall
You don't give any hint of your platform (pc? smartphone? by hand? In automated program?)
But as a great starting point working with a few points by hand, in maps.google.com if you put the latitude and longitude (may be longitude and latitude try both) it will show them to you on the map. Google maps even has simple APIs so you can use this functionality reasonably simply (for example from a matlab program).
A client wants a page on their website where the user can search for a stockist of their product range. What they want is the ability to enter a UK PostCode and for a list of their stockists to appear in closest order.
Is there a way to utilise Google Maps to determine the closest supplier?
I have a PHP MySQL database with ALL the suppliers Postcodes and details and I'll have the postcode of the user to use as well.
I think you'll need the Ordnance Survey CodePoint data to map your postcodes to a longitude and latitude. Codepoint is available as a commercial product, or there's also Codepoint Open
Once you have the latitude and longitude, then you can use standard algorithms such as Haversine or Vincenty to calculate distances between the points
EDIT
Structure of the CodePoint Open CSV files can be found here. Note that location is held as Northings and Esatings rather than Longitude and Latitude, so it will need converting. There's a number of articles about this on the web, e.g.
http://www.deepbluesky.com/blog/-/converting-os-coodinates-into-longitude-latitude_7/
http://mediakey.dk/~cc/convert-northing-and-easting-utm-to-longitude-and-latitude/
but you need to be aware that OS Northings/Eastings are based on the Airy 1830 ellipsoid rather than the WGS84 model used by Google maps (and most GSM systems). Failing to allow for this difference can put you out by anything between 70-120m between Cornwall and East Anglia.
You can also find PHP functions to do this conversion at the Movable Type site (essential reading for any PHP developer, working with GeoData. I'd recommend adding a couple off columns to your stockist database for longitude and latitude, and a one-off script to update all existing data using the Codepoint data, then modify your insert/update routines for stockists to keep this information up-to-date.
Using PHP, another solution for this conversion is PHPCoord by Jonathan Stott
For higher performance in your database query, do a lookup of lat/long against a "bounding box" before calculating distances. I've explained how to do this in response to a similar query.
I think you'll need a little more data than that. I'd suggest that you need to store the latitude and longitude of every UK postcode alongside the postcode itself. I think this data is available from the Royal Mail for a cost and I remember reading somewhere that it was going to be available in the public domain too. Have a user enter their postcode, look up their latitude and longitude in the db and then use that to perform another query that calculates the closest supplier to them, perhaps within a certain number of miles. You could perhaps create a stores procedure to do all of this on the db. This post seems to have some details on how to do this.
If you want it to be fast then pre-compute the distances between postcodes within a certain radius of each other - see the following:
Calculate distance between zip codes and users
This should work fine in your case unless you think a user would be willing to travel more than 100 miles to purchase a product ??
Instead of Google Maps, this open postcode geocoding API may be of use to you. As others have mentioned, once you have lat long for both points you can use standard algos to find distances. A previous question contains some info on how to do this directly in SQL.
You may find this website helpful: SQL/CSV Postcode Database
I've previously used this to lookup long/lat from a postcode "outcode".
There's also a php script (which I've not tested) which calculates distance between two postcodes.
Google maps has a function that lets you retreive the distance between two points: getDistance(), of the GDirections class.
Now, this is Javascript, so it'll only be able to calculate once the page has loaded.
My search form however, has the ability to indicate what's the maximum distance you want between yourself and another person and based on that, a list of search results has to be provided.
Since I can't use this function in PHP, is there another way for me to calculate the distance between 2 points on the earth? By giving up the street, postal code and city name, just like what Google maps needs.
This page has a list of google maps parameters for use in http connection. By specifying the output paramter you can choose to give back kml files or similar and can be used in any lanuage that can make http connections.
Looks like you want to calculate a Great Circle Distance
Formulas have been discussed here on stackoverflow before.
A point on earth is defined by it's latitude and longitude. If you want to calculate the distance between 2 points on earth by giving up the street, postal code and city name, you will need geo-referene data.
This data is available for free on the internet, but the accuracy and availability differ greatly from region to region. (USA data is of good quality, data for Kenia for example will be harder to come by)
So to answer your question:
Yes, there are other ways to calculate what you want. But they require more work/are more complex than just querying the google API.
You might want tot read: Creating a Store Locator with PHP, MySQL & Google Maps
Hope this points you into the right direction.
You can use kml file. It's xml-formatted file that you can recieve by link like http://maps.google.com/maps?f=d&hl=en&saddr=<src latitude>,<src longitude>&daddr=<dist latitude>,<dist longitude>&ie=UTF8&0&om=0&output=kml
in recieved file you can parse and summate distances from