R help - Aug 2004 - (Lattice) How to improve the readability of a bwplot, i.e. separating groups somehow

Hi all,
first of all thanks for the answer to my previous question on lattice some
time ago.
In particular to Patrick Connolly for advices on netiquette (I hope this
time I'm doing right....)
and to Deepayan Sarkar fro the help on lattice.

Now, my nowaday problem.

Please consider the following

mydf<-cbind.data.frame(
RESPONSE = c(rnorm(9,rep(2:4,each=3),10),
             rnorm(9,rep(7:9,each=3),10),
             rnorm(9,rep(5:7,each=3),10),
             rnorm(9,rep(9:11,each=3),10)
             ),
STD =   c(rep("Std.Ext",18), rep("Std.Int",18)),
METHOD =  rep(c("A","B"),2, each=9),
VIAL =  rep(paste("Vial",rep(1:3, each=3)),4))
library(lattice)
my.theme<-list(background=list(col=0),
               strip.background=list(col="transparent"),
               box.dot=list(cex=0.1, col=1),
               box.umbrella=list(col=1,lty=1),
               box.rectangle=list(col= 1))
lset(my.theme)
bwplot(paste(METHOD,VIAL)~RESPONSE|STD, data=mydf)

as a (fictitious) experiment on a determination of a substance in 3 vials,
which was quantified
with an external (or internal) standard,
with two different methods (A or B)
with 3 injections per vial (replicates)

I would like to stress the difference between A and B (Method) in the
bwplot,
so I imagine I could distantiate the boxplots, or colour them according to
the "Method".

How can I add a blank line between "A Vial 3" and "B Vial
1"?
Or how can I change the colour of the boxplots? (say the three lower ones in
black, and the rest red)

Thanks

Ottorino-Luca Pantani, Universit?? di Firenze
Dip. Scienza del Suolo e Nutrizione della Pianta
P.zle Cascine 28 50144 Firenze Italia
Tel 39 055 3288 384 Fax 39 055 333 273 OLPantani at unifi.it
http://www.unifi.it/dssnp/

On Tuesday 03 August 2004 06:39, 8rino-Luca Pantani
wrote:
> Hi all,
> first of all thanks for the answer to my previous question on lattice
> some time ago.
> In particular to Patrick Connolly for advices on netiquette (I hope
> this time I'm doing right....)
> and to Deepayan Sarkar fro the help on lattice.
>
> Now, my nowaday problem.
>
> Please consider the following
>
> mydf<-cbind.data.frame(
> RESPONSE = c(rnorm(9,rep(2:4,each=3),10),
>              rnorm(9,rep(7:9,each=3),10),
>              rnorm(9,rep(5:7,each=3),10),
>              rnorm(9,rep(9:11,each=3),10)
>              ),
> STD =   c(rep("Std.Ext",18), rep("Std.Int",18)),
> METHOD =  rep(c("A","B"),2, each=9),
> VIAL =  rep(paste("Vial",rep(1:3, each=3)),4))
> library(lattice)
> my.theme<-list(background=list(col=0),
>                strip.background=list(col="transparent"),
>                box.dot=list(cex=0.1, col=1),
>                box.umbrella=list(col=1,lty=1),
>                box.rectangle=list(col= 1))
> lset(my.theme)
> bwplot(paste(METHOD,VIAL)~RESPONSE|STD, data=mydf)
>
> as a (fictitious) experiment on a determination of a substance in 3
> vials, which was quantified
> with an external (or internal) standard,
> with two different methods (A or B)
> with 3 injections per vial (replicates)
>
> I would like to stress the difference between A and B (Method) in the
> bwplot,
> so I imagine I could distantiate the boxplots, or colour them
> according to the "Method".
Why not

bwplot(VIAL ~ RESPONSE | STD + METHOD, data=mydf)

?
> How can I add a blank line between "A Vial 3" and "B Vial
1"?
You could create a factor with a fake blank level. e.g.,


mydf$METHOD.VIAL <-
    factor(with(mydf, paste(METHOD, VIAL)), 
           levels = c("A Vial 1", "A Vial 2", "A Vial
3",
                      " ", "B Vial 1", "B Vial 2",
                      "B Vial 3"))

bwplot(METHOD.VIAL ~ RESPONSE | STD, data=mydf, drop = FALSE)

> Or how can I change the colour of the boxplots? (say the three lower
> ones in black, and the rest red)
That's going to be non-trivial. If you are really desparate, look at 

http://tolstoy.newcastle.edu.au/R/help/04/02/0848.html

Hope that helps,

Deepayan