[Maptitude] WinBugs

Thu, 29 Jun 2006 12:28:46 -0700

Oh yes I have and I have Jon Wakefield, a substantial WinBugs guru, just up the road at UW. I have some philosophical problems will full Bayesian rate determination. Using a non-informative prior "bugs" me. Empirical Bayes seems OK as one does start with a distribution that makes more sense in spite of it being from the data itself. But I might get over my philosophical objections. I tend to use all the tools and see what happens. And WinBugs seems to go along with the whole crowd.
 
Since this is Maptitude list, I would suggest with regard to folks doing a lot of work (like WinBugs, GeoDA, SaTScan, and many other spatial stat apps) that to support all these uses that Caliper develop a bridge with a major stat package. ESRI has done that with SAS but it is unbelievelably expensive, especially for what you get. My choice would be STATA as it is pretty easy to use compared to SAS, and has excellent graphics. I suppose one could  build a bridge  in GISDK, and as long as one did not try to tonally integrate STATA into Maptitude, it might be relatively simple. One would have a drop down menu item with the basic items that would allow Stata to connect to a selected dataview. Or import a STATA file, not just record level data, but statistical output results.
 
I do this anyway by exporting dataviews as DBFs and then read into STATA through StatTransfer and back to a dbf for Maptitude. But having all that go a lot more quickly would be very useful.
 
Dick H

From: [email protected] [mailto:[EMAIL PROTECTED] On Behalf Of Aniruddha Banerjee
Sent: Thursday, June 29, 2006 10:47 AM
To: [email protected]
Subject: Re: [Maptitude] Question about the lurking legend

Hi Richard,
 
Have you tried WinBUGs software that uses conditional hierarchical models (Bayesian)? It's the most elegant method of doing spatial analysis (having mathematical rigor and at the same time quite simple...).
 
Rudy B.
Berkeley, CA.

 
On 6/29/06, Hoskins, Richard E. (DOH) <[EMAIL PROTECTED]> wrote:

I am looking at the spatial determinants of low birth weight in the
urban counties of W Washington. We have superb birth certificate data
which I have geocoded. I am particularly interested in finding clusters
of LBW with SaTScan (www.satscan.org) and then comparing the SES of the
clusters with RR > 1 with those < 1. So far there are no surprises RR>1
are way poorer, more minorities, etc. All the usual stuff. But the
result is valuable in that we need to know exactly where the big
problems are for developing interventions. Of course LBW is distributed
through out the area, but it does cluster in 4 or 5 areas. I have not
compared these areas to the location of WIC clinics. One thing I am
interested in is developing a disparity index that works for our state.
So I am looking at clustering techniques (Fastclus in SAS, k-means) and
seeing what I come up with. ESRI gave me their Tapestry, market
segmentation data to play with. It is usually horribly expensive so out
of the reach of public health practitioners. When I finally got the
details from them about how they developed their categories (65 of
them!) I realized that they are not state specific enough to be useful.
Low SES in WA is different than low SES in NY, especially in urban
settings. I think I can develop a spate specific index which others can
do as well from census data, land value assessments, etc..

Dick H



-----Original Message-----
From: [email protected] [mailto: [email protected]] On
Behalf Of Susan Kinne
Sent: Wednesday, June 28, 2006 3:26 PM
To: [email protected]
Subject: [Maptitude] Question about the lurking legend

When I do a quick-and-dirty copy and paste of a map into a Word file, I
end up with two legends--one where I left it, attached to the map, and
the other in the opposite upper corner. I can't figure how to get rid
of the extra legend (nor have I figured out exactly how to open the
legend in a separate window when I detach it from the map).

Help would be appreciated. I can't find anything about this in the
manual.

Susan Kinne

>

------------------------ Yahoo! Groups Sponsor --------------------~-->
Yahoo! Groups gets a make over. See the new email design.
http://us.click.yahoo.com/XISQkA/lOaOAA/yQLSAA/C5grlB/TM
----------------------------------------------------------~->

Yahoo! Groups Links


__._,_.___


SPONSORED LINKS
Geographic information systems Geography game Geography

YAHOO! GROUPS LINKS



__,_._,___

Reply via email to