Re: calculating distance based on zip/postal codes

by Jan Theodore Galkowski <jtgalkowski(at)alum.mit.edu>

 Date:  Mon, 15 May 2000 08:05:06 -0400
 To:  Susan Duncan <sduncan(at)malico.com>
 Cc:  hwg-business <hwg-business(at)hwg.org>
 In-Reply-To:  malico
  todo: View Thread, Original
At 11:45 PM 05/14/2000 -0400, Susan Duncan wrote:
>I have an application whereby I need to calculate the distance
>between two points using zip/postal codes.  Other sites out there
>do it, so there must be a way of handling this short of typing in
>a slew of codes and taking measurements.

Well, once ZIP or Postal codes are translated to geographic=20
coordinates, the calculation is straightforward as long as one
is satisfied with results from a spherical model of Earth.  A
spherical model is pretty good since Earth is very close to a
sphere, about 3 feet out of round (up/down) in 17 miles along its=20
circumference.  If one is comparing to GPS-grade positions,
however, there's no choice but to use the oblate spheroid
model of Earth which is slightly elliptical seen on a cross-
section through the poles.  There are smaller effects which=20
pile up in places:  On a ship, the tide elevation at center
of the oceans is a significant effect, and there are Earth
tides, too, on the order of centimeters.

So, if you're interested in distances between places, I'd use
the Zip Code to latitude, longitude conversion, and calculate
those distances.  When you get to calculating distances between
geographic points, don't use the arc-cosine rule:  It requires
notoriously high precisions for closer and closer points.

  --jtg

[snip]



______________________________________________________________________
 Jan Theodore Galkowski   =B0o=B0   (:-)}        demiourgos(at)smalltalk.org=20
 algebraist.com/       www.marsociety.org/   jtgalkowski(at)alum.mit.edu=20
 PGP fingerprint: 2757 F86D AA51 677D 38D7  964B 9A8D 7852 A494 3790=20

HTML: hwg-business mailing list archives, maintained by Webmasters @ IWA