Skip to content

boxplot in subgroups

3 messages · Pooja Jain, Annette Heisswolf

Pooja Jain
# 
Hi,

I have data that looks like this:

ASA1    ASA2    C1_C2
C       M       9.0225
S       S       2.4315
M       C       3.4894
M       S       4.5282
C       M       1.3183
C       S       1.3735
S       C       1.0488
S       M       7.948
M       C       4.5827

I need to plot Boxplots for a given ASA1 (either C,S, or M) with  
respect to C1_C2. However, instead of one boxplot I want to plot three  
boxplots for a given ASA1 value such that each sub-boxplot represent  
the distribution for C1-C2 for each of the three possible values of  
ASA2 (C,S or M). I want to show this subgroup with a single xtics  
(either C, S or M) corresponding to the ASA1 value selected.

I tried to do it, but ending with the following error. I am not sure  
how exactly I should build the object to plot as subgroup.


Error in axis(side = 1, at = 1:9, labels = c("C", "M", "S" :
   'at' and 'labels' lengths differ, 9 != 3
In addition: Warning message:
In bxp(list(stats = c(1e-04, 2.82745, 6.0193, 10.5957, 22.2476,  :
   some notches went outside hinges ('box'): maybe set notch=FALSE


I would greatly appreciate any help in this matter.

Thank you.


-Kanu

This message has been checked for viruses but the contents of an attachment
may still contain software viruses, which could damage your computer system:
you are advised to perform your own checks. Email communications with the
University of Nottingham may be monitored as permitted by UK legislation.
Annette Heisswolf
# 
Hei,

the error message below just tells you that you have specified nine places along the axis where you want to have labels (i.e. at=1:9) but 
obviously you have only specified three labels (i.e. labels = c("C","M","S")). Thus, R is complaining that there are not enough labels for 
all positions. However, I guess from your description that you want to have only three labels below your plot, one for each group of three 
boxes that share the same ASA1 value. Thus, you could for example do like this:

axis(side=1,at=c(2,5,8),labels=c("C","S","M"))

Then you would get the label always below the middle one of the three boxes of each group.

Now the boxes are, however, still distributed evenly across the plot, but in case you want to have those that belong to the same subgroup a 
bit closer together, you can specify their position via the "at" option on the boxplot function (see ?boxplot).

An example with random data might look like this:

ASA1=factor(rep(c("C","S","M"),each=100))
ASA2=factor(rep(c("C","S","M"),100))
C1_C2=runif(300)
boxplot(C1_C2~ASA2*ASA1,xlim=c(0,12),at=c(1:3,5:7,9:11),xaxt="n")
axis(side=1,at=c(2,6,10),labels=c("C","S","M"))

See ?axis for further options, e.g. tcl=0 in case you don't want to have a tick at the axis at the position of the label.

Hope that helps,

Annette

Pooja Jain schrieb:

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Pooja Jain
    

    

      
      #
    

  


  

      
Thank you very much Annette,

With your help I can do exactly what I wanted to.

Best wishes,
-Pooja

        
On 26 Mar 2009, at 06:59, Annette Heisswolf wrote: