Here are 4 for a start:
1. MI users need a definite link to a statistical package. Even better, a
stat package that has spatial routines. S-Plus by Insightful Software,
Seattle has a link to AV and with the S-Plus Spatial Stats routine. Its is a
bit bumpy, but it works. When you fire up AV you get a set of menus for
Splus and its spatial analysis routines. There are features that allow
determination of nearest neighbors, etc. Full accounting for spatial
autocorrelation in regression routines. ESRI is not all that interested in
S-Plus, and GIS vendors say there is not really a market. Not clear S-Plus
developers will push Spatial Stats because of ESRI's lack of interest. I
know they are wrong. The business community needs spatial stat analysis at
least as much as academics do. S-Plus is a good choice because of the
fantastic graphics they offer.
MI could also use as a base SpaceStat by Luc Anslen www.spacestat.com Its
still a DOS program but VERY powerful. It cries out for a real GIS
connection and a GUI interface.
Spatial stats as a methodology has arrived, but it appears the GIS community
(other than IDRISI and manifold) is not interested. Big mistake.
2. Along the lines of above, but a way to cross-tab records (like Proc Freq
in SAS, Crosstabs in SPSS)
3. Empirical Bayes smoothing - allows accounting for small number problems
which arise in calculating rates for covariates in large scale maps. I can
provide a reference list. A method that can allow business types to assess
whether rates (whether disease rates or per capita income or unemployment or
crime rates) as calculated by ordinary methods really reflect what is really
going on.
4. Aerial interpolation. Buffers or other polygons overlayed over a target
layer to estimate the attributes of one layer from another. I have
population and incomes levels in a set of block groups, how do I estimate
those numbers for a school district which cuts across block group
boundaries, or for a set of buffers around, say, a cell phone tower or a
toxic waste site? This cannot be done in any vector desktop GIS except
Maptitude www.caliper.com (that I am aware of) It can be done in ArcInfo
but not without pulling a tooth first.
R Hoskins
WA State Dept of Health
Olympia, WA USA
GMT -8
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]]On Behalf Of Karpovich,
Robert
Sent: Tuesday, June 19, 2001 10:17 AM
To: 'Mapinfo-L (E-mail)
Subject: MI-L MapInfo improvements
I've been a MapInfo reseller since the first Windows version. Each year a
attend the reseller conference and the question is asked what improvements
are needed in MIPro. The issue of the new version of the ESRI product and
the release of MIPro 6.5 has caused some comments on MapInfo-L. We all know
that some desired features don't have the benefit versus cost appeal. The
folks at MapInfo read Mapinfo-L. So here is the question. What is missing
and what is needed to move the product forward? I'll print out the
responses and if I go to the annual meeting I will give a copy to the MI Pro
folks. In any event, the MapInfo folks will see your responses. The MI Pro
developers are good folks and do listen. They have the same desire and
drive I see on Mapinfo-L. Certain needed improvements may stick out to the
MIPro developers when you respond.
Bob Karpovich
UAI Environmental
_______________________________________________________________________
List hosting provided by Directions Magazine | www.directionsmag.com |
To unsubscribe, send e-mail to [EMAIL PROTECTED] and
put "unsubscribe MapInfo-L" in the message body.
_______________________________________________________________________
List hosting provided by Directions Magazine | www.directionsmag.com |
To unsubscribe, send e-mail to [EMAIL PROTECTED] and
put "unsubscribe MapInfo-L" in the message body.
