From rhoskins@home.com Thu Jun 24 17:47:58 1999 Received: from mxu3.u.washington.edu (mxu3.u.washington.edu [140.142.33.7]) by lists.u.washington.edu (8.9.3+UW99.02/8.9.3+UW99.01) with ESMTP id RAA48186 for ; Thu, 24 Jun 1999 17:47:57 -0700 Received: from ha1.rdc1.wa.home.com (siteadm@ha1.rdc1.wa.home.com [24.0.2.66]) by mxu3.u.washington.edu (8.9.3+UW99.02/8.9.3+UW99.06) with ESMTP id RAA02695 for ; Thu, 24 Jun 1999 17:47:56 -0700 Received: from c501552a ([24.5.121.123]) by ha1.rdc1.wa.home.com (Netscape Mail Server v2.02) with SMTP id AAA17804 for ; Thu, 24 Jun 1999 17:47:55 -0700 Message-ID: <003701bebea4$5bac8720$7b790518@olmpi1.wa.home.com> From: "Richard E. Hoskins" To: References: <3.0.32.19990624152805.00ffc40c@iris.caliper.com> Subject: Re: ZIP lat/longs Date: Thu, 24 Jun 1999 17:47:53 -0700 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2615.200 X-MimeOLE: Produced By Microsoft MimeOLE V5.00.2615.200 Someone on our list gave us a site where free centroids are available and I think for years other than 1990. I seem to have lost the URL. If you are trying to get some distance measure, centroids are OK but you'd better examine a map and see that you are calculating something meaningful. One approach is to borrow or get someone with the Zips for the year you need to generate centroids. Most GIS programs will do that. If I had say, a list of cancer cases and wanted to geocode them and subsequently calculate disease rates or similar, I'd be looking to get the cases assigned to a Zipcode that has a base population for the denominator. If some of my cases has a post office only Zipcode, then I would have to assign a Zipcode to the address that was an enclosing Zipcode of the post office only or point Zipped. If I had a Zipped and not a long/lat then the assignment could be done randomly using a weighted function of the population densities of the surrounding Zipcodes. One approach as far as calculating inter-Zipcode distances is to use the centroid of the most populous census tract in the Zipcode or better yet, a weighted centroid of the underlying census tracts (block groups are better) where the weighting factor is the population. Dick Hoskins .