R help - Nov 2009 - Set colors in a PCA plot based on a gradient vector

Hi all,

 

I'm making a PCA plot with eight variables (columns taken from a larger data
frame "fieldTrial0809[idx,c(39:46)]"). I then want the symbols in the
plot to be colored as a gradient from red to blue, depending on the value of
another column in "fieldTrial0809[idx, c(48)]" containg temperatures
from -12.1 to -5.4.

 

I don't want to use the heat.colors(n) function, because then I just get
colors from red to yellow...

 

I've created a vector "colTemperature" with 12 colors and I've
also made a new vector "tempACC", based on the values of
"fieldTrial0809[idx, c(48)]" with 12 levels.

 

What I don't know, is how to incorporate this vector in the text() function
below:

> df=fieldTrial0809[idx,c(39:46)]
> PC=prcomp (df[,1:8])
> hcl=hclust(dist(df))
> plot(PC$x[,1],PC$x[,2], col="white", main="8
variables")
> text(PC$x[,1],PC$x[,2], labels=cutree(hcl,8), col=cutree(hcl,8))   # Change
this, but how???
>
> col1=rgb(0.164,0.043,0.850)   # Set the colors
> col2=rgb(0.150,0.306,1.000)
> col3=rgb(0.250,0.630,1.000)
> col4=rgb(0.450,0.853,1.000)
> col5=rgb(0.670,0.973,1.000)
> col6=rgb(0.880,1.000,1.000)
> col7=rgb(1.000,1.000,0.750)
> col8=rgb(1.000,0.880,0.600)
> col9=rgb(1.000,0.679,0.450)
> col10=rgb(0.970,0.430,0.370)
> col11=rgb(0.850,0.150,0.196)
> col12=rgb(0.650,0.000,0.130)
>
colTemperature=c(col1,col2,col3,col4,col5,col6,col7,col8,col9,col10,col11,col12)
# Put all colors in a vector
> 
> tempLT50_ACC=fieldTrial0809[idx, c(48)]
> tempACC=cut(tempLT50_ACC, breaks=c(-5.4, -5.958333333, -6.516666667,
-7.075, -7.633333333, -8.191666667, -8.75, -9.308333333, -9.866666667, -10.425,
-10.98333333, -11.54166667,-12.1))   # Define the breaks
> table(tempACC)
tempACC
(-12.1,-11.5]   (-11.5,-11]   (-11,-10.4] (-10.4,-9.87] (-9.87,-9.31]
(-9.31,-8.75]
           25            21            24            21           131           
28
(-8.75,-8.19] (-8.19,-7.63] (-7.63,-7.08] (-7.08,-6.52] (-6.52,-5.96] 
(-5.96,-5.4]
           33            14            11             7            17           
11

 

Finally, I've got another question, how do you decide how many components to
use in "text(... labels=cutree(hcl,8), ...)"? Do you try with 
different values until i looks nice or what..?

 

Best regards,

 

Joel
 		 	   		  
_________________________________________________________________
Hitta kärleken i vinter!
http://dejting.se.msn.com/channel/index.aspx?trackingid=1002952
	[[alternative HTML version deleted]]

On 11/06/2009 03:21 AM, Joel F?rstenberg-H?gg wrote:
> Hi all,
>
>
>
> I'm making a PCA plot with eight variables (columns taken from a larger
data frame "fieldTrial0809[idx,c(39:46)]"). I then want the symbols in
the plot to be colored as a gradient from red to blue, depending on the value of
another column in "fieldTrial0809[idx, c(48)]" containg temperatures
from -12.1 to -5.4.
>
>
>
> I don't want to use the heat.colors(n) function, because then I just
get colors from red to yellow...
>
>
>
> I've created a vector "colTemperature" with 12 colors and
I've also made a new vector "tempACC", based on the values of
"fieldTrial0809[idx, c(48)]" with 12 levels.
>
>
>
> What I don't know, is how to incorporate this vector in the text()
function below:
>
>
>    
>> df=fieldTrial0809[idx,c(39:46)]
>> PC=prcomp (df[,1:8])
>> hcl=hclust(dist(df))
>>      
>    
>> plot(PC$x[,1],PC$x[,2], col="white", main="8
variables")
>> text(PC$x[,1],PC$x[,2], labels=cutree(hcl,8), col=cutree(hcl,8))   #
Change this, but how???
>>      
library(plotrix)
text(...,col=color.scale(fieldTrial0809[idx,48],
  extremes=c("red","blue"))
> Finally, I've got another question, how do you decide how many
components to use in "text(... labels=cutree(hcl,8), ...)"? Do you try
with  different values until i looks nice or what..?
>    
>
As in so many other areas of life, beauty is in the eye of the beholder.

Jim