From rhoskins@home.com Mon Jun 21 15:50:03 1999 Received: from mxu2.u.washington.edu (mxu2.u.washington.edu [140.142.32.9]) by lists.u.washington.edu (8.9.3+UW99.02/8.9.3+UW99.01) with ESMTP id PAA11360 for ; Mon, 21 Jun 1999 15:50:02 -0700 Received: from ha1.rdc1.wa.home.com (siteadm@ha1.rdc1.wa.home.com [24.0.2.66]) by mxu2.u.washington.edu (8.9.3+UW99.02/8.9.3+UW99.06) with ESMTP id PAA22805 for ; Mon, 21 Jun 1999 15:50:02 -0700 Received: from c501552a ([24.5.121.123]) by ha1.rdc1.wa.home.com (Netscape Mail Server v2.02) with SMTP id AAA23473 for ; Mon, 21 Jun 1999 15:50:01 -0700 Message-ID: <011d01bebc38$62184220$7b790518@olmpi1.wa.home.com> From: "Richard E. Hoskins" To: References: <199906212202.PAA15968@mxu2.u.washington.edu> Subject: Re: Free Zipcodes? Date: Mon, 21 Jun 1999 15:49:55 -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 Alice: You have special considerations in GISing health data that we in the lower 48 do not have. I'd be curious to hear your special concerns and issues. I'll bet some of them would touch home with the rest of us, especially for others in areas where the population density varies widely. (when you think about, most everywhere) As far as the Zipcode issue goes, it's time for public health types to lobby for useful geocoding of health data. As someone suggested, likely the Post Office could help (rural routes) and the rest left to the TIGER files (the US street files) to effectively geocode health data. Public Health people rightly worry about privacy issues, but I'll bet the marketing and credit report people, not to mention the IRS (the US tax people) have more data on where we live, etc. than public health people could ever imagine having for improving the Nation's health. If public health people could get at least a close idea where the sick people are and where the exposures are we could forget about problems with Zipcodes and the like. Dick Hoskins ----- Original Message ----- From: Rarig, Alice To: Sent: Monday, June 21, 1999 2:56 PM Subject: RE: Free Zipcodes? > I think you are wise to be so cautious -- it all depends on the part of the > country you are in, and the particulars of the postal delivery, what you can > say about the population distribution within the zipcode. Here in Alaska we > have some zipcodes with vast empty spaces, and a household or two at some > remote point... > > -----Original Message----- > From: Richard E. Hoskins [mailto:rhoskins@home.com] > Sent: Monday, June 21, 1999 1:18 PM > To: waphgis@u.washington.edu > Subject: Re: Free Zipcodes? > > > I have tried a weighting scheme internal to the Zipcode using the population > densities of the underlying census tracts. It makes for cool looking maps > but one is imputing stuff and aggregating in the wrong direction. I think it > still might be OK if one could come up with a scheme to compute an estimate > for the variance. So if I had an age adjusted rate for prostate cancer at > the Zip level, the weighted scheme could estimate what the rate would be at > the underlying census tract level, but I do not know how to get at the > variance. Statistical territory with crevasses I think. > > Dick Hoskins > > > ----- Original Message ----- > From: Andy Amir-Fazli > To: > Sent: Monday, June 21, 1999 1:23 PM > Subject: Re: Free Zipcodes? > > > > At 09:26 PM 6/19/99 -0700, you wrote: > > >On some reflection, I think the health/GIS community needs to forget > about > > >post/Zipcodes as a useful unit of spatial analysis. > > > > Thanks Dick for bringing up an important topic. I agree Zip codes are a > > lousy way to aggregate or position trend data. However, this is the only > > tool we really have to assign locations universally when geocoding. In NM > > as in WA (and elsewhere) many people use PO Boxes, RRs, local roads > > known only to locals, and street addresses not known to Tiger, GDT, Etak, > > etc. The zip code at least gives us a rough proximity to home or work. > > > > I'm convinced many people live in NM so they cannot be located by govt > > officials, health or otherwise! :) However, everybody likes to get mail. > > > > The key issue is how to conflate a zip code location to census tract or > > population based spatial unit. I've been working on weighting zip code > > points to population centers (rural NM is very clustered) and I've seen > > someone else try to use a gravity model. Has anyone solved this problem? > > > > >Down with zip/post codes ... > > > > I say put zipcodes in their place, as a filtering tool to meaningful > > "health districts" or census areas. > > > > Regards, > > Andy > > ============================================================ > > Andy Amir-Fazli - Database Administrator > > Epidemiology and Cancer Control > > New Mexico Tumor Registry (National Cancer Institute) > > University of New Mexico Phone: 505.272.8575 > > 2325 Camino de Salud NE Fax: 505.272.8572 > > Albuquerque, NM 87131-5306 aamirf@nmtr.unm.edu > > > > ============================================================ > > .