LISTSERV at the University of Georgia
Menubar Imagemap
Home Browse Manage Request Manuals Register
Previous messageNext messagePrevious in topicNext in topicPrevious by same authorNext by same authorPrevious page (May 2006, week 4)Back to main SAS-L pageJoin or leave SAS-L (or change settings)ReplyPost a new messageSearchProportional fontNon-proportional font
Date:         Wed, 24 May 2006 13:23:36 -0400
Reply-To:     Kitty Lee <lee.kitty@YAHOO.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Kitty Lee <lee.kitty@YAHOO.COM>
Subject:      Re: LAT/LONG COORDINATES--DISTANCE BETWEEN THEM

I have seen studies which show that the correlation between air distance (any straight-line distance calculated with a math formula) and traffic distance is LOW. One study done in Ohio suggests the correlation is only around .2.

Depends on your research question---how important is the measure of distance? Do you really need to measure it precisely/as accurate as possible?

I mean, if you're doing an accessibility study then I would say you definitely need to use traffic distance. But if you're just using distance as a proxy measure of something (e.g. feeling of 'closeness'), or you're using distance as 'ranks' (i.e. you don't care about the actual distance, and what you care is the relative distance --- A is closer to B and is closer to C). If you're using distance to find relative ranking of places, then I'd say air distance may be okay. It also depends on the areal scale of your study. If you're measuring distance between Los Angeles and Reno, then traffic distance would usually agree with air distance.

So it all depends on your research question.

If you have access to GIS (ArcMap, ArcInfo and Network Analyst). I can show you how to turn streetmaps into network and find traffic distance between points. If you have many points, essentially you can ask for a distance matrix among all the points. :-)

K.

On Tue, 23 May 2006 15:44:57 -0500, Nick . <ni14@MAIL.COM> wrote:

>Yes, TRAVEL DISTANCE by car is more accurate for me. So, given two points in a city, I know the LAT/LONG of these two points, how can I find the distance by car between them? If one point in ZIP_COD1 has LAT1/LONG1 and the other point in ZIP_CODE2 has LAT2/LONG2, how can I approximate the car distance between them? Well, I don't think anyone can come up with a mathematical formula which will measure the turns and twists of the road. All we can do, I suppose, is find the straight line between them. I can't use MapQuest. That's not the spirit of my question. If I only had one point, yes, I suppose I could do it. But if I have many points, can I do it? >


Back to: Top of message | Previous page | Main SAS-L page