Choosing the most accurate geocode

Question

Am I using http://maps.google.com/maps/geo ? web service for geocoding some addresses.

The problem is that a more complete address does not necessarily give a more accurate geocode.

for example, in Llantisilio, Denbyshire, UK is much more accurate than Llantysilio, Llangollen, Denbighshire, UK

The Accuracy attribute in XML is not very helpful in choosing the address to select.

How did other people deal with this problem? Is there a good way to choose the best geocode that works mostly / all the time?

* edit A bit of additional information - when I put a more complete address, the first line of the address is ignored, and the geocoder goes to another, but exact address, which is the central street located in the additional line added to the address. In this example, he selects Castle Street on average llangollen, apparently ignoring Llantysilio.

Edit kdgregory: here are two API requests I used (missing an API key, it seems no problem):

http://maps.google.com/maps/geo?q=Llantysilio,+Denbighshire,+UK&sensor=false&output=xml http://maps.google.com/maps/geo?q=Llantysilio,Llangollen,++Denbighshire,+UK&sensor=false&output=xml 

Answer 1

In conclusion, I came to the conclusion that in the UK there are too many strange errors in the consistency of addresses with google gesoding geocoding in the UK, but in the end I managed to figure out a way to use postal codes, which is much more accurate: how is this done

Answer 2

You must interpret the accuracy of my friend. As a rule, 2 parts have accuracy, first an address game. An important role is played by the second part. You can geocode something at a precision level in the United States or at a city level, centralized zip code, interpolated street level, or actual parcel accuracy level. The first example has 4, and the second is 9. For this service, above.

Accuracy Value Description 0 Unknown accuracy. 1 Accuracy at country level. 2 Region (state, province, prefecture, etc.). 3 The accuracy of the level of the subregion (district, municipality, etc.). 4 The accuracy of the city (city, village). 5 Zip code level accuracy (zip code). 6 Street level stability. 7 The accuracy of the level of intersection. 8 Address accuracy. 9 Premises (the name of the building, the name of the property, a shopping center, etc.).

Answer 3

It may be good to note that Google does not comply with the XAL specifications , but rather implements them in a subset.

So this means that you may not be able to:

 place.AddressDetails.Country.AdministrativeArea.Locality.LocalityName place.AddressDetails.Country.AdministrativeArea.AdministrativeAreaName place.AddressDetails.Country.CountryName 

As the country and subregion can be provided, but the administrative area - not.

The returned data is identified using an accuracy scale that gives you a relative idea of ​​what you can expect from the data. Thus, you can store objects and tear off parts of the full address using this variable and try to geocode in this way - not recommended.

As a rule, a full address (without travel) is a good way to find a common location. You can use some weighted preferential logic that Google provides to refine the address.

eg. Use setViewPort or setCountryCode to give your search a little more precision.

Remember that geocoding is not a science . You cannot expect consistent results.

Answer 4

The answer to the geocode. The symbol [0] through gmap you can check what you have and take a level or try again. I chose the default value in

 place.AddressDetails.Country.AdministrativeArea.Locality.LocalityName place.AddressDetails.Country.AdministrativeArea.AdministrativeAreaName place.AddressDetails.Country.CountryName 

This can be more logically named as shown above. gmaps 3 works somewhat incompatible with v2.

Answer 5

You can try a very ugly hack that consists in geocoding your full address and all the subsets of words that your address contains, you get a lot of geocodes that you use to get the links associated with them using the reverse geocoding tool. When you have a lot of addresses, you compare them with the one you first gave, then you take the most accurate geocode ... Many requests, many iterations grow with every word that you add to your address, and ugly work, but it can be interesting to do some statistics ^^