R help - Mar 2011 - Spatial cluster analysis of continous outcome variable

Dear R Users, R Core
Team,
I have a two dimensional space where I measure a numerical value in two
situations at different points. I have measured the change and I would like to
test if there are areas in this 2D-space where there is a different amount of
change (no change, increase, decrease). I don´t know if it´s better to analyse
the data just with the coordinates or if its better to group them in
"pixels" (and obtain the mean value for each pixel) and then run the
cluster analysis. I would like to know if there is a package/function that
allows me to do these calculations.I would also like to know if it could be done
in a 3D space (I have collapsed the data to 2D because I don´t have many points.
Thanks in advance
J Toledo 		 	   		  
	[[alternative HTML version deleted]]

Did you post your data or hypothetical data?
Usually that helps make your problem more clear and more interesting
( likely to get a useful response to your post).



----------------------------------------
From: tintin_jb at hotmail.com
To: r-help at r-project.org
Date: Thu, 17 Mar 2011 17:38:14 +0100
Subject: [R] Spatial cluster analysis of continous outcome variable



Dear R Users, R Core
Team,
I have a two dimensional space where I measure a numerical value in two
situations at different points. I have measured the change and I would like to
test if there are areas in this 2D-space where there is a different amount of
change (no change, increase, decrease). I don?t know if it?s better to analyse
the data just with the coordinates or if its better to group them in
"pixels" (and obtain the mean value for each pixel) and then run the
cluster analysis. I would like to know if there is a package/function that
allows me to do these calculations.I would also like to know if it could be done
in a 3D space (I have collapsed the data to 2D because I don?t have many points.
Thanks in advance





J Toledo
        [[alternative HTML version deleted]]

I attach the data (csv format). There are the 3 coordinates, (but as there are
not so many points I wanted two do 3 analysis in each of them collapsing one
variable).There are two variables to study I have posted the data as a ratio
between both states and as a percentage state between both states. The data are
from different samples (and each sample has 3 or 6 measures).Thanks again.
> From: marchywka at hotmail.com
> To: tintin_jb at hotmail.com; r-help at r-project.org
> Subject: RE: [R] Spatial cluster analysis of continous outcome variable
> Date: Thu, 17 Mar 2011 15:20:09 -0400
> 
> 
> 
> 
> 
> 
> 
> 
> Did you post your data or hypothetical data?
> Usually that helps make your problem more clear and more interesting
> ( likely to get a useful response to your post).
> 
> 
> 
> ----------------------------------------
> From: tintin_jb at hotmail.com
> To: r-help at r-project.org
> Date: Thu, 17 Mar 2011 17:38:14 +0100
> Subject: [R] Spatial cluster analysis of continous outcome variable
> 
> 
> 
> Dear R Users, R Core
> Team,
> I have a two dimensional space where I measure a numerical value in two
situations at different points. I have measured the change and I would like to
test if there are areas in this 2D-space where there is a different amount of
change (no change, increase, decrease). I don?t know if it?s better to analyse
the data just with the coordinates or if its better to group them in
"pixels" (and obtain the mean value for each pixel) and then run the
cluster analysis. I would like to know if there is a package/function that
allows me to do these calculations.I would also like to know if it could be done
in a 3D space (I have collapsed the data to 2D because I don?t have many points.
> Thanks in advance
> 
> 
> 
> 
> 
> J Toledo
>         [[alternative HTML version deleted]]
> 
>

I read it but it said PDF file and Ps, didn´t specify which other files, so I
attached a csv file, which I thought would work.I have uploaded the file in
rapidhare (second option was putting on the
web):http://rapidshare.com/files/453101614/Coordinates_and_values.csv
Hope this works.Thanks for your interest.
> From: dwinsemius@comcast.net
> To: tintin_jb@hotmail.com
> Subject: Re: [R] Spatial cluster analysis of continous outcome variable
> Date: Thu, 17 Mar 2011 22:12:01 -0400
> 
> No data came through. You probably didn't read the Posting Guide or if
> you did you didn't read it closely enough.
> 
> -- 
> David.
> On Mar 17, 2011, at 4:11 PM, Jon Toledo wrote:
> 
> >
> > I attach the data (csv format). There are the 3 coordinates, (but as  
> > there are not so many points I wanted two do 3 analysis in each of  
> > them collapsing one variable).There are two variables to study I  
> > have posted the data as a ratio between both states and as a  
> > percentage state between both states. The data are from different  
> > samples (and each sample has 3 or 6 measures).Thanks again.
> >
> >> From: marchywka@hotmail.com
> >> To: tintin_jb@hotmail.com; r-help@r-project.org
> >> Subject: RE: [R] Spatial cluster analysis of continous outcome  
> >> variable
> >> Date: Thu, 17 Mar 2011 15:20:09 -0400
> >>
> >> Did you post your data or hypothetical data?
> >> Usually that helps make your problem more clear and more
interesting
> >> ( likely to get a useful response to your post).
> >>
> >>
> >>
> >> ----------------------------------------
> >> From: tintin_jb@hotmail.com
> >> To: r-help@r-project.org
> >> Date: Thu, 17 Mar 2011 17:38:14 +0100
> >> Subject: [R] Spatial cluster analysis of continous outcome
variable
> >>
> >>
> >>
> >> Dear R Users, R Core
> >> Team,
> >> I have a two dimensional space where I measure a numerical value
in
> >> two situations at different points. I have measured the change and
> >> I would like to test if there are areas in this 2D-space where  
> >> there is a different amount of change (no change, increase,  
> >> decrease). I don´t know if it´s better to analyse the data just  
> >> with the coordinates or if its better to group them in  
> >> "pixels" (and obtain the mean value for each pixel) and
then run
> >> the cluster analysis. I would like to know if there is a package/ 
> >> function that allows me to do these calculations.I would also like
> >> to know if it could be done in a 3D space (I have collapsed the  
> >> data to 2D because I don´t have many points.
> >> Thanks in advance
> >>
> >>
> >>
> >>
> >>
> >> J Toledo
> >>        [[alternative HTML version deleted]]
 		 	   		  
	[[alternative HTML version deleted]]

Hi everyone,

Is there any command to identify the pattern of responses of a database 
with this format:

year  id
2008	1
2009	1
2008	2
2009	2
2008	3
2009	3
2008	4
2009	4
2010	4


I just need the frequency of the patterns grouped by id:

2008 2009 2010 = 80
         2009 2010 = 30
2008 2009         = 10
and so on....

Thank you in advance!

-- 
Sebasti?n Daza
sebastian.daza at gmail.com

Finally, I solved my problem using the following procedure and a 
database called ej

ej$a <- 1

head(ej)

    ano nunico a
1 2008      1 1
2 2009      1 1
3 2008      2 1
4 2009      2 1
5 2008      3 1
6 2009      3 1

library(reshape)
dej <- cast(ej, nunico ~ ano, sum, margins = FALSE)

head(dej)

   nunico 2008 2009 2010
1      1    1    1    0
2      2    1    1    0
3      3    1    1    0
4      4    1    1    1
5      5    1    0    1
6      6    1    1    1

dej[dej==0] <- NA

library(Hmisc)
na.pattern(dej[,c(2:4)])


pattern
  000  001  010  011  110
3385 1073  203  338  573

That way I can review the pattern of my panel data.

Sebastian