Geocoding is the process of estimating a latitude
and longitude for a given location. Geocoder.us is
a public service providing
free geocoding of addresses and intersections in
the United States. The geocoding service relies on
Geo::Coder::US,
a Perl module available for download from the
CPAN.
The US Census Bureau provides location data for nearly
every street and highway in the US in the form of the
TIGER/Line
data set, which runs to over 24 gigabytes
uncompressed. Geo::Coder::US condenses the 2004 TIGER/Line
data set down to a 750 megabyte BerkeleyDB database. A given
address or intersection is parsed into its components, and the
matching address range is searched for in the database. The
address is then interpolated between the endpoints of the
matching address range to estimate its latitude and longitude.
For urban addresses, the coordinates given by Geo::Coder::US
can be within 10 or 15 meters of the actual location --
which is well within the range of typical GPS accuracy!
Other times the coordinates can be off, so use your own
judgement when using the data!
From the
Tiger Metadata
"NAD83 is the coordinate datum used for the 48 contiguous
states, the District of Columbia, Alaska, Puerto Rico, and
the Virgin Islands of the United States. Regional datums are
used for Hawaii and the Pacific Island Areas."
And
"The TIGER/Line data are not in a mapping projection even
though most of the features were scanned directly from
source maps (usually USGS 1:100,000-scale topographic quads)
that were in a projection."
Read the metadata yourself to get a better understanding of
what this means for your applications, and to get a feel for the
magnitude of the task that the Census did (rah, rah, go census!).
You bet it is! Commercial US geocoding services can
charge up to $1 per lookup, but the TIGER/Line data on
which their geocoders (and ours) are based is in the public
domain. The Geo::Coder::US suite is made available free for
download
under either the
GPL
v2 or the
Perl
Artistic License. All data from the free services
on the geocoder.us website are provided under a
NonCommercial-ShareAlike
license, which means you can use anything you get from
us for any non-commercial purpose, so long as you give us
proper credit and extend the same courtesy to others.