Geocoding via Geowanking

| | Comments (0) | TrackBacks (0)

This software detects street addresses (such as "2400 Bayshore Parkway,
Mountain View, CA 94043") in a corpus of text and converts them into
geographical coordinates (such as "122.09720 W, 37.42532 N"). These
coordinates are indexed in a two-dimensional index along with a conventional
keyword index of the corpus. A query processor is then able to rapidly
process queries which ask for documents which match certain keywords and/or
contain addresses within a certain radius of a specified target address.
(Think "bookstores near me".)

Sonny Parafina reminds me that the winner of Google's programming contest last year is this amazingly versatile and useful Google geocoding app by Dan Egnor. Details here. And download here.

via Geowanking, my new favorite list. Big signal, little noise.

The problem, of course, is that this kind of data isn't publicly available for other countries, thus making it difficult to bridge the gap between addresses that are useful to normal people (like mailing addresses) and addresses that are useful to computers (like lat/long).

0 TrackBacks

Listed below are links to blogs that reference this entry: Geocoding via Geowanking.

TrackBack URL for this entry: http://www.confectious.net/mt/mt-tb.cgi/7

Leave a comment

Pages

  • /thinking
  • projects

Archives