R help - Jul 2011 - superimposing different plot types in lattice panel.superpose

I would like to plot 3 best-fit models in a single panel of a lattice plot,
superimposed on 3 corresponding datasets in the same panel.  My goal is to
show the models as lines of 3 different colors, and the data as points whose
colors correspond to the model colors.  In essence, I have two levels of
grouping: 1) model vs. data, and 2) model number.  Since there is only one
?groups? variable, I have tried to deal with the additional grouping level
by subsetting the data inside a custom panel function (basically a hack),
but something about the way parameters are passed in panel.superpose (I
think) is making it hard to show both points and lines. 

My question is very similar to a previous post:
http://www.ask.com/web?q=r%20panel.superpose%20bwplot%20sim%20actual&o=15527&l=dis&prt=NIS&chn=retail&geo=US&ver=18
, but the questioner in that case was using bwplot, which automatically
makes a separate plot for every level of the categorical variable, so they
didn?t face the two-level grouping problem, and I have been unable to figure
out how to adapt their answer.  Another approach I tried was to put my model
function inside of my custom panel function so that the analysis occurred
there, but I couldn?t get it to subset the x and y data appropriately.

In the toy problem below I want to plot each inverted V (?model?) as LINES,
with a single POINT (?data?) in the center of each inverted V, the same
color as the inverted V.  The code runs, but I can?t seem to mix lines and
points.  In my real problem (stable isotopes with ellipses superimposed on
data) there will be additional panels but I am creating only one panel here,
for simplicity.

#generate test "model results"
x<-1:9
y<-rep(c(1,3,1),3)
model<-c(rep("a",3),rep("b",3),rep("c",3))
modelresults<-data.frame(x,y,model)

#generate test "data"
x<-c(2,5,8)
y<-rep(1.5,3)
model<-c("a","b","c")
data<-data.frame(x,y,model)

#combine them into one data set
combined<-make.groups(modelresults,data)

#custom panel function
panel.dualplot <- function(...) { 
   
panel.xyplot(x[which="modelresults"],y[which="modelresults"],...)
    panel.points(x[which="data"],y[which="data"],...)
} 

#main call to xyplot
xyplot(y ~ x, data=combined,
type = "l",
panel = panel.superpose,
groups = model,
panel.groups = panel.dualplot,
)


I?d be very grateful for any suggestions.  

Thanks!

John



--
View this message in context:
http://r.789695.n4.nabble.com/superimposing-different-plot-types-in-lattice-panel-superpose-tp3642808p3642808.html
Sent from the R help mailing list archive at Nabble.com.

Hi:

Here's one way out, if I read your intention properly...

plot1 <- xyplot(y ~ x, data = modelresults, groups = model, type =
'l',
                        xlim = c(0, 10), ylim = c(0.5, 3.5))
plot2 <- xyplot(y ~ x, data = data, groups = model, type = 'p',
                        pch = 16, xlim = c(0, 10), ylim = c(0.5, 3.5))
plot1 + plot2

You might need latticeExtra to combine the plots (??); I had it loaded
with lattice when I did the plot.

HTH,
Dennis

On Sun, Jul 3, 2011 at 6:52 PM, genghis <jcpayne at uw.edu>
wrote:
> I would like to plot 3 best-fit models in a single panel of a lattice plot,
> superimposed on 3 corresponding datasets in the same panel. ?My goal is to
> show the models as lines of 3 different colors, and the data as points
whose
> colors correspond to the model colors. ?In essence, I have two levels of
> grouping: 1) model vs. data, and 2) model number. ?Since there is only one
> ?groups? variable, I have tried to deal with the additional grouping level
> by subsetting the data inside a custom panel function (basically a hack),
> but something about the way parameters are passed in panel.superpose (I
> think) is making it hard to show both points and lines.
>
> My question is very similar to a previous post:
>
http://www.ask.com/web?q=r%20panel.superpose%20bwplot%20sim%20actual&o=15527&l=dis&prt=NIS&chn=retail&geo=US&ver=18
> , but the questioner in that case was using bwplot, which automatically
> makes a separate plot for every level of the categorical variable, so they
> didn?t face the two-level grouping problem, and I have been unable to
figure
> out how to adapt their answer. ?Another approach I tried was to put my
model
> function inside of my custom panel function so that the analysis occurred
> there, but I couldn?t get it to subset the x and y data appropriately.
>
> In the toy problem below I want to plot each inverted V (?model?) as LINES,
> with a single POINT (?data?) in the center of each inverted V, the same
> color as the inverted V. ?The code runs, but I can?t seem to mix lines and
> points. ?In my real problem (stable isotopes with ellipses superimposed on
> data) there will be additional panels but I am creating only one panel
here,
> for simplicity.
>
> #generate test "model results"
> x<-1:9
> y<-rep(c(1,3,1),3)
> model<-c(rep("a",3),rep("b",3),rep("c",3))
> modelresults<-data.frame(x,y,model)
>
> #generate test "data"
> x<-c(2,5,8)
> y<-rep(1.5,3)
> model<-c("a","b","c")
> data<-data.frame(x,y,model)
>
> #combine them into one data set
> combined<-make.groups(modelresults,data)
>
> #custom panel function
> panel.dualplot <- function(...) {
> ?
?panel.xyplot(x[which="modelresults"],y[which="modelresults"],...)
> ? ?panel.points(x[which="data"],y[which="data"],...)
> }
>
> #main call to xyplot
> xyplot(y ~ x, data=combined,
> type = "l",
> panel = panel.superpose,
> groups = model,
> panel.groups = panel.dualplot,
> )
>
>
> I?d be very grateful for any suggestions.
>
> Thanks!
>
> John
>
>
>
> --
> View this message in context:
http://r.789695.n4.nabble.com/superimposing-different-plot-types-in-lattice-panel-superpose-tp3642808p3642808.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at r-project.org 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.
>