[R] how to add 3d-points to bplot {rms} figure?

[Bert Gunter]

Below.

-- Bert

On Mon, Apr 9, 2012 at 7:39 PM, David Winsemius <dwinsemius at comcast.net> wrote:
>
> On Apr 9, 2012, at 6:24 PM, phi771 wrote:
>
>> Hello!
>> I have created a bplot-figure using this code:
>>
>> *file <- "2dcali_red.ttt"
>> ux<-as.matrix(read.table(file, dec = ","))
>>
>> mode(ux)<-'numeric'
>>
>> vel<-ux[,1]
>> ang<-ux[,2]
>> x<-ux[,3]
>> y<-ux[,4]
>>
>> dat<- data.frame(ang=ang, x=x,y=y)
>>
>> require(rms)
>>
>> ddist2 <- datadist(dat)
>> options(datadist="ddist2")
>>
>> fitn <- lrm(ang ~ rcs(x,4) + rcs(y,4), data=dat)
>> predi <- Predict(fitn, x, y)
>>
>> bplot(predi,lfun=wireframe, screen = list(z = -40, x = -80), drape=TRUE)*
>>
>> The file 2dcali_red.ttt  consists of 4 columns can be found here :
>> http://www.color-space.de/upload/dl/2dcali_red.ttt link
>>
>> The code gives me the bplot-figure, which looks like this:
>>
>> http://r.789695.n4.nabble.com/file/n4544050/Screen_shot_2012-04-10_at_12.04.24_AM.png
>>
>> Now I want to add some 3d data (x,y,z) to this plot (like scatterplot3d),
>> but i don't know how. I also don't know how to extract the x,y,z mesh as
>> coordinates from the bplot. Some help would be very appreciated. Thank
>> you!
>
>
> That is most probably a lattice::wireframe object. You could use
> 'trellis.focus' or the 'layer' functions in pkg::latticeExtra to attempt
> panel.3dbars() additions. I haven't actually done this, but I have added
> panel function results to the lattice::contourplots that are created
> byrms/Hmisc.
>
> The 3d aspect adds a significant level of extra work:
And is generally a bad idea. Perspective plots of all types are much
over-hyped. They are pretty, but the very act of creating artificial
perspective for non-realistic shapes can distort perception and lead
to misinterpretation of data. Parts of the scene hide other parts; it
is difficult to quantitatively judge depths of valleys and heights of
hills; etc. And, as David notes, it is very difficult to get "right."

> http://tolstoy.newcastle.edu.au/R/e2/help/06/10/3274.html
>
> Maybe a contourplot would suffice?
-- and is generally a much better, albeit less sexy, idea.

>
> --
> David Winsemius, MD
> West Hartford, CT
>
> ______________________________________________
> 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.



-- 

Bert Gunter
Genentech Nonclinical Biostatistics

Internal Contact Info:
Phone: 467-7374
Website:
http://pharmadevelopment.roche.com/index/pdb/pdb-functional-groups/pdb-biostatistics/pdb-ncb-home.htm