R help - Sep 2006 - Weighted association map

Could somebody program this kind of plot type to R, if none exists,  
based on mds or correlation tables or some more suitable method? What  
do you think about idea? Does it work? None similar or better exists?

http://weightedassociationmap.blogspot.com/


Atte Tenkanen
University of Turku, Finland

> Could somebody program this kind of plot type to R, if none exists,  
> based on mds or correlation tables or some more suitable method?  
> What do you think about idea? Does it work? None similar or better  
> exists?
>
> http://weightedassociationmap.blogspot.com/
>
>
> Atte Tenkanen
> University of Turku, Finland


I got a hint of package ade4 and function dudi.pca. Something like  
this, though I can't immediately interpret those dudi.pca-pictures. I  
originally saw this as a more general idea, which can, if wanted, be  
applied with statistical methods too. If you do multidimensional  
scaling, you can see, how "near" all components are each other, but  
you can't see the connections or directions. If you have a  
correlation matrix, you can't see the connections graphically.

Atte

Hi R-users,
I am using stripchart with coincident points,

y <-rbinom(100, 3, 0.5)
stripchart(y, method="stack", pch="o", vertical=TRUE)

But the result is not centered in the sense that if a value (say value 0) is 
repeated 7 times, the first point is ploted in the middle and the rest at 
its right side, in stead of ploting 3 at its right and 3 at its left.  Can 
anybody help?

Thanks,
Berta.

Try the sna package.  Below we calculate the
correlation matrix, kor, of the numeric cols of builtin iris
dataset.  Zap negative ones and discretize rest to
get lwd width matrix, lwd, used for edge widths.  From
that create the adjacency matrix, sign(lwd), and plot it
using indicated layout mode.  Seems like three of
the variables are correlated and Sepal.Width is uncorrelated
or negatively correlated to those. Try playing around with
gplot args to create variations.

library(sna)
set.seed(123) # layout uses random numbers
kor <- cor(iris[1:4])
lwd <- replace(kor, TRUE, 10 * round(pmax(0, kor), 1))
gplot(sign(lwd), edge.lwd = lwd, displaylabels = TRUE, label = rownames(kor))

On 9/8/06, kone <attenka at utu.fi> wrote:
> Could somebody program this kind of plot type to R, if none exists,
> based on mds or correlation tables or some more suitable method? What
> do you think about idea? Does it work? None similar or better exists?
>
> http://weightedassociationmap.blogspot.com/
>
>
> Atte Tenkanen
> University of Turku, Finland
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

Actually the discretization does not appear to be needed.  This
works just as well:

set.seed(123)
kor <- cor(iris[1:4])
gplot(sign(kor), edge.lwd = 10*kor, displaylabels = TRUE, label = rownames(kor))

On 9/8/06, Gabor Grothendieck <ggrothendieck at gmail.com>
wrote:
> Try the sna package.  Below we calculate the
> correlation matrix, kor, of the numeric cols of builtin iris
> dataset.  Zap negative ones and discretize rest to
> get lwd width matrix, lwd, used for edge widths.  From
> that create the adjacency matrix, sign(lwd), and plot it
> using indicated layout mode.  Seems like three of
> the variables are correlated and Sepal.Width is uncorrelated
> or negatively correlated to those. Try playing around with
> gplot args to create variations.
>
> library(sna)
> set.seed(123) # layout uses random numbers
> kor <- cor(iris[1:4])
> lwd <- replace(kor, TRUE, 10 * round(pmax(0, kor), 1))
> gplot(sign(lwd), edge.lwd = lwd, displaylabels = TRUE, label =
rownames(kor))
>
> On 9/8/06, kone <attenka at utu.fi> wrote:
> > Could somebody program this kind of plot type to R, if none exists,
> > based on mds or correlation tables or some more suitable method? What
> > do you think about idea? Does it work? None similar or better exists?
> >
> > http://weightedassociationmap.blogspot.com/
> >
> >
> > Atte Tenkanen
> > University of Turku, Finland
> >
> > ______________________________________________
> > R-help at stat.math.ethz.ch mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> > and provide commented, minimal, self-contained, reproducible code.
> >
>

You may be looking for:
 stripchart(y, method="jitter", pch="o",
vertical=TRUE,jitter=.5)

Rob
____________________________
Robert W. Baer, Ph.D.
Associate Professor
Department of Physiology
A. T. Still University of Health Science
800 W. Jefferson St.
Kirksville, MO 63501-1497 USA

----- Original Message ----- 
From: "Berta" <ibanez at bioef.org>
To: <r-help at stat.math.ethz.ch>
Sent: Friday, September 08, 2006 3:53 AM
Subject: [R] how to construct stripchart with coincident points centered

> Hi R-users,
> I am using stripchart with coincident points,
>
> y <-rbinom(100, 3, 0.5)
> stripchart(y, method="stack", pch="o", vertical=TRUE)
>
> But the result is not centered in the sense that if a value (say value 0)
is
> repeated 7 times, the first point is ploted in the middle and the rest at
> its right side, in stead of ploting 3 at its right and 3 at its left.  Can
> anybody help?
>
> Thanks,
> Berta.
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>

Sorry to answer so late, but this is just what I want ;-)

-Atte

Gabor Grothendieck kirjoitti 8.9.2006 kello 14.53:
> Actually the discretization does not appear to be needed.  This
> works just as well:
>
> set.seed(123)
> kor <- cor(iris[1:4])
> gplot(sign(kor), edge.lwd = 10*kor, displaylabels = TRUE, label =  
> rownames(kor))
>
> On 9/8/06, Gabor Grothendieck <ggrothendieck at gmail.com> wrote:
>> Try the sna package.  Below we calculate the
>> correlation matrix, kor, of the numeric cols of builtin iris
>> dataset.  Zap negative ones and discretize rest to
>> get lwd width matrix, lwd, used for edge widths.  From
>> that create the adjacency matrix, sign(lwd), and plot it
>> using indicated layout mode.  Seems like three of
>> the variables are correlated and Sepal.Width is uncorrelated
>> or negatively correlated to those. Try playing around with
>> gplot args to create variations.
>>
>> library(sna)
>> set.seed(123) # layout uses random numbers
>> kor <- cor(iris[1:4])
>> lwd <- replace(kor, TRUE, 10 * round(pmax(0, kor), 1))
>> gplot(sign(lwd), edge.lwd = lwd, displaylabels = TRUE, label =  
>> rownames(kor))
>>
>> On 9/8/06, kone <attenka at utu.fi> wrote:
>> > Could somebody program this kind of plot type to R, if none
exists,
>> > based on mds or correlation tables or some more suitable method?  
>> What
>> > do you think about idea? Does it work? None similar or better  
>> exists?
>> >
>> > http://weightedassociationmap.blogspot.com/
>> >
>> >
>> > Atte Tenkanen
>> > University of Turku, Finland
>> >
>> > ______________________________________________
>> > R-help at stat.math.ethz.ch mailing list
>> > https://stat.ethz.ch/mailman/listinfo/r-help
>> > PLEASE do read the posting guide http://www.R-project.org/ 
>> posting-guide.html
>> > and provide commented, minimal, self-contained, reproducible code.
>> >
>>

Thanks Rob for the clue, it is nearly  but not exactly what I am looking 
for.
If we plot the first figure as I want (but not centered) and the second 
figure as you suggest:

 win.graph(); par(mfrow=c(2,1));  y <-rbinom(100, 3, 0.5)
 stripchart(y, method="stack", pch="o", vertical=TRUE);
abline(v=1)
 stripchart(y, method="jitter", pch="o",
vertical=TRUE,jitter=.5);
abline(v=1)

we obtain "more or less" a symmetric plot in the second figure but
points
within the same value are not equally distanced as I wish (sorry, I did not 
tell this in the previous message). Is there a method in which I can obtain 
the first figure but with coincident points centered and equally distanced?

Thanks again,

Berta.
> You may be looking for:
> stripchart(y, method="jitter", pch="o",
vertical=TRUE,jitter=.5)
>
> Rob
> ----- Original Message ----- 
>> Hi R-users,
>> I am using stripchart with coincident points,
>>
>> y <-rbinom(100, 3, 0.5)
>> stripchart(y, method="stack", pch="o",
vertical=TRUE)
>>
>> But the result is not centered in the sense that if a value (say value
0)
> is
>> repeated 7 times, the first point is ploted in the middle and the rest
at
>> its right side, in stead of ploting 3 at its right and 3 at its left. 
>> Can
>> anybody help?
>>
>> Thanks,
>> Berta.

Great.  In looking at it once more I realize that even the sign(kor)
part could be shortened to just kor in this example.  Regards.

library(sna)
set.seed(123)
kor <- cor(iris[1:4])
gplot(kor, edge.lwd = 10*kor, displaylabels = TRUE, label = rownames(kor))



On 9/11/06, kone <attenka at utu.fi> wrote:
> Sorry to answer so late, but this is just what I want ;-)
>
> -Atte
>
> Gabor Grothendieck kirjoitti 8.9.2006 kello 14.53:
>
> > Actually the discretization does not appear to be needed.  This
> > works just as well:
> >
> > set.seed(123)
> > kor <- cor(iris[1:4])
> > gplot(sign(kor), edge.lwd = 10*kor, displaylabels = TRUE, label >
> rownames(kor))
> >
> > On 9/8/06, Gabor Grothendieck <ggrothendieck at gmail.com>
wrote:
> >> Try the sna package.  Below we calculate the
> >> correlation matrix, kor, of the numeric cols of builtin iris
> >> dataset.  Zap negative ones and discretize rest to
> >> get lwd width matrix, lwd, used for edge widths.  From
> >> that create the adjacency matrix, sign(lwd), and plot it
> >> using indicated layout mode.  Seems like three of
> >> the variables are correlated and Sepal.Width is uncorrelated
> >> or negatively correlated to those. Try playing around with
> >> gplot args to create variations.
> >>
> >> library(sna)
> >> set.seed(123) # layout uses random numbers
> >> kor <- cor(iris[1:4])
> >> lwd <- replace(kor, TRUE, 10 * round(pmax(0, kor), 1))
> >> gplot(sign(lwd), edge.lwd = lwd, displaylabels = TRUE, label >
>> rownames(kor))
> >>
> >> On 9/8/06, kone <attenka at utu.fi> wrote:
> >> > Could somebody program this kind of plot type to R, if none
exists,
> >> > based on mds or correlation tables or some more suitable
method?
> >> What
> >> > do you think about idea? Does it work? None similar or better
> >> exists?
> >> >
> >> > http://weightedassociationmap.blogspot.com/
> >> >
> >> >
> >> > Atte Tenkanen
> >> > University of Turku, Finland
> >> >
> >> > ______________________________________________
> >> > R-help at stat.math.ethz.ch mailing list
> >> > https://stat.ethz.ch/mailman/listinfo/r-help
> >> > PLEASE do read the posting guide http://www.R-project.org/
> >> posting-guide.html
> >> > and provide commented, minimal, self-contained, reproducible
code.
> >> >
> >>
>
>