Re: [R] lattice 3x3 plot: force common y-limits accross rows and align x-axes

[Boris Vasiliev]

Duncan and Bert,

Thank you very much for your help with my question.  It's very much 
appreciated.


I used your suggestions to get the plot I needed: 

* 3x3 lattice of dotplots, 

* x-limits are the same for all panels, 

* y-limits and y-ticks in each row are the same, 

* y-limits and y-ticks are different between rows,  

* within each row subjects are ordered by sum(count) in the row.  


My code is below just in case somebody in the r-help list finds it useful.

Regards,
Boris.


# raw data

df <- data.frame(
        subject=c('A','A','A',
                  'BB','BB',
                  'CCC','CCC','CCC',
                  'DD','DD',
                  'A','A','A',
                  'FFFF','FFFF',
                  'A','A',
                  'B','B'),
        risk=c('high','high','high',
               'high','high',
               'high','high','high',
               'med','med',
               'med','med','med',
               'med','med',
               'low','low',
               'low','low'),
        treatment=c('none','optX','optZ',
                    'none','optZ',
                    'none','optX','optZ',
                    'none','optZ',
                    'none','optX','optZ',
                    'none','optZ',
                    'none','optX',
                    'none','optZ'),
        count=c(5,10,2,
                3,5,
                8,1,2,
                3,7,
                10,2,5,
                15,2,
                7,7,

                10,8))


# re-level factors
df$risk <- factor(df$risk,levels=c('low','med','high'))
df$treatment <- factor(df$treatment,levels=c('none','optX','optZ'))


# create unique subjects ordered by sum(count) and risk
df$sbj.byrisk <- paste(df$risk,df$subject,sep="_")
df$sbj <- reorder(reorder(df$sbj.byrisk,df$count,sum),as.numeric(df$risk))
df$sbj.ix <- as.numeric(df$sbj)

# for each row (i.e. risk), find limits, ticks, labels, and
# maximum number of points per panel
df.byrisk <- split(df,df['risk'])
ylim <- lapply(df.byrisk,function(idf) return(range(idf$sbj.ix)))
ytck <- lapply(df.byrisk,function(idf) return(sort(unique(idf$sbj.ix))))
ylbl <- lapply(df.byrisk,function(idf) {
                          ytck <- 
sort(unique(idf$sbj.ix))
                          jdnx <- 
pmatch(ytck,idf$sbj.ix)
                          ylbl <- 
sub(".*_","",idf$sbj[jdnx])
                          return(ylbl)
                        })
yhei <- unlist(lapply(ytck,length))

# set up lists for limits, ticks, labels, panel heights
ylims <- rep(ylim,c(3,3,3))
ytcks <- rep(list(NULL,NULL,NULL),c(3,3,3))
ytcks[seq(1,7,by=3)] <- ytck
ylbls <- rep(list(NULL,NULL,NULL),c(3,3,3))
ylbls[seq(1,7,by=3)] <- ylbl

# set up plot layout
laywid <- list(axis.panel=c(1,0.1,0.1))
layhei <- list(panel=yhei/sum(yhei))

# plot
oltc <- dotplot(sbj~count|treatment+risk,data=df,
                type=c("p","h"),origin=0,
                scales=list(x=list(limits=c(-1,16),
                                   
alternating=FALSE),
                            y=list(relation="free",
                                   
alternating=FALSE,
                                   
limits=ylims,
                                   at=ytcks,
                                   
labels=ylbls)),
                par.settings=list(layout.widths=laywid,
                                  
layout.heights=layhei),
                between=list(y=0.2))
useOuterStrips(oltc)


________________________________
 From: Duncan Mackay <mac...@northnet.com.au>

oject.org> 
Sent: Monday, February 18, 2013 10:29:23 PM
Subject: Re: [R] lattice 3x3 plot: force common y-limits accross rows  and 
align x-axes

Hi Boris

Just a different take on it, quick glimpse of it

library(lattice)
library(latticeExtra)

z= unique(df$subject)
df$nsub <- sapply(df$subject, pmatch, z)

useOuterStrips(xyplot(count ~ nsub|treatment*risk, df, type = "h", lwd = 2, 
scales = list(x = list(at = 1:6, labels = z))))

The order can be changed by changing the order in z by making subject a factor 
in the correct order

see

http://finzi.psych.upenn.edu/R/Rhelp02/archive/43626.html

on how to change the yaxis limits. An example of mine (cut and paste)  for an 
8x3 panel xyplot (it could be streamlined - i needed to change things as I went 
along to get everything in)

           par.settings =  layout.heights = list(panel = 
c(1,1,0.6,0.6,0.6,0.6,1,0.5)/sum(c(1,1,0.6,0.6,1,0.6,1,0.5)) )
                          ),
           scales   = list(x = list(alternating = FALSE,
                                    relation    = "same",
                                    rot         = 0),
                           y = list(alternating = FALSE,
                                    relation    = "free",
                                    limits = 
list(c(-1500,30000),c(-1500,30000),c(-1500,30000),  #  BZ
                                                  
c(-1500,30000),c(-1500,30000),c(-1500,30000), #  LV
                                                  
c(-2000,20000),c(-2000,20000),c(-2000,20000), #  L2
                                                  c( 
-500, 4000),c( -500, 4000),c( -500, 4000),  #  CL
                                                  
c(-2000,20000),c(-2000,20000),c(-2000,20000), #  OP
                                                  
c(-2000,20000),c(-2000,20000),c(-2000,20000), #  AB
                                                  
c(-1500,30000),c(-1500,30000),c(-1500,30000), #  MO
                                                  
c(-500, 4000),c(-500,  4000),c(-500,  4000)), #  MX
                                    at = 
rep(list(seq(0,30000,10000), NULL,
                                                  
seq(0,30000,10000), NULL,
                                                  
seq(0,20000,10000), NULL,
                                                  
seq(0, 4000, 2000), NULL,
                                                  
seq(0,20000,10000), NULL,
                                                  
seq(0,20000,10000), NULL,
                                                  
seq(0,30000,10000), NULL,
                                                  
seq(0, 4000, 2000), NULL ),
                                                   
c(1,2, 1,2, 1,2, 1,2, 1,2,  1,2, 1,2, 1,2) ),
                                    labels = 
rep(list(as.character(seq(0,30,10)), NULL,
                                                     
 as.character(seq(0,30,10)), NULL,
                                                     
 as.character(seq(0,20,10)), NULL,
                                                     
 as.character(seq(0, 4, 2)), NULL,
                                                     
 as.character(seq(0,20,10)), NULL,
                                                     
 as.character(seq(0,20,10)), NULL,
                                                     
 as.character(seq(0,30,10)), NULL,
                                                     
 as.character(seq(0, 4, 2)), NULL ),
                                                     
 c(1,2, 1,2, 1,2, 1,2, 1,2,  1,2, 1,2, 1,2) ),
                                    rot         = 0)
                      ),

I had different panel heights for each row  so if they were all the same it 
would be

panel = rep(1,n)/n

the trick is to get the rep of the lists to match what would normally happen 
for each of the all the panels ie whether an axis/label is required

str(plot object) can help if problems

HTH

Duncan

Duncan Mackay
Department of Agronomy and Soil Science
University of New England
Armidale NSW 2351
Email: home: mac...@northnet.com.au



At 12:22 19/02/2013, you wrote:
> Content-Type: text/plain
> Content-Disposition: inline
> Content-length: 14523
> 
> Bert,
> 
> Thank you very much for your reply.  Your suggestions to set manually 
> y-limits of panels in lattice plot certainly puts me on the right track.
> 
> It works really well with xyplot():
> 
> hf <- data.frame(
>    x=seq(1,20,by=1),
>    y=seq(0.1,2,by=0.1),
>    risk=gl(2,10,labels=c("low","high")),
>    treatment=rep(gl(2,5,labels=c("oXX","oYY")),2))       
> Â Â Â Â Â Â
> Â
> xyplot(y~x|treatment+risk,data=hf,type="b",
>        scales=list(y=list(limits=list(c(-0.1,1.1),c(-0.1,1.1),
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â 
>             c(1.0,2.1),c(1.0,2.1)),
>                    
> at=list(seq(0,1,0.2),NULL,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â 
>   seq(1,2,0.2),NULL),
>                    relation="free")))
> 
> With the xyplot() above, I get 2x2 lattice where panels in the same row have 
> the same y-limits:
> 
> Panel 1,2 (row 1): -0.1, 1.1
> Panel 3,4 (row 2): 1.0, 2.1
> 
> Unfortunately, it does not work so well with dot plots:
> 
> 
> kf <- data.frame(subject=c("A","BB","A","C","DDD","DDD"),
>                  
> risk=c("low","low","low","high","high","high"),
>                  
> treatment=c("oXX","oXX","oYY","oXX","oXX","oYY"),
>                  count=c(5,7,10,4,13,8))
> kf$risk <- factor(kf$risk,levels=c("low","high"))
> kf$treatment <- factor(kf$treatment,levels=c("oXX","oYY"))
> 
> lims.low <- factor(c("A","BB"))
> lims.high <- factor(c("C","DDD"))
> dotplot(subject~count|treatment+risk,data=kf,type=c("p","h"),
>         scales=c(y=list(limits=list(lims.low,lims.low,
> Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â Â 
>      lims.high,lims.high),
>                         
> relation="free")))
> 
> 
> With the dotplot() above I get 2x2 lattice plot where all panels have  the 
> same y-limits; y-ticks on all panels are 'A','BB','C','DDD'.
> 
> 
> This behaviour is probably fixed by design.  According to documentation 
> for prepanel functions: "... to make this information consistent between 
> panels, the ‘xlim’ or ‘ylim’ values should 
> represent all the levels of the corresponding factor, even if some are not 
> used within that particular panel."
> 
> A way to trick lattice would be to use xyplot() with a custom panel 
> function.  I'll give a try today or tomorrow.
> 
> Regards,
> Boris.
> 
> 
> 
> 
> ________________________________
>  From: Bert Gunter <gunter.ber...@gene.com>
> 
> Cc: "mac...@northnet.com.au" <mac...@northnet.com.au>; "r-help@r-project.org" 
> <r-help@r-project.org>
> Sent: Monday, February 18, 2013 1:09:10 PM
> Subject: Re: [R] lattice 3x3 plot: force common y-limits accross rows and 
> align x-axes
> 
> Boris:
> 
> If I understand you correctly, you wish to set panel limits by row. I
> know of no "slick" way of doing this (others may), as what is in a row
> can change depending on layout and how you determine the scale for the
> rows' contents may depend on the application context.
> 
> So I would do it manually: First, outside of trellis, compute your
> xlim and ylim values for each row as you desire. Then issue the
> trellis call, with the relation="free" component in the scales = list,
> Then just use the list form of the xlim and ylim argument to xyplot
> with your precomputed limits.
> 
> -- Bert
> 
> 
> > Hi Duncan,
> >
> > Thank you for quick reply.
> >
> > I am not sure that your solution solves the problem.  If I use
> > useOuterStrips(dotplot(count ~ subject|risk*treatment,df))
> > the order of subjects and panel y-limits in each panel are
> >
> > A, B, BB, CCC, DD, FFFF.
> >
> > However, the order of subjects and y-limits which I would
> > like to get are
> >
> > Panels 1,2,3 (row 1): A, B
> >
> > Panels 4,5,6 (row 2): DD, A, FFFF
> >
> > Panels 7,8,9 (row 3): BB, CCC, A.
> >
> > I can re-index subjects so that they have correct order in each row.
> > However, I am not sure how to specify per-panel y-limits in lattice
> > so that all panels in the same row have the same y-limits irrespective
> > of what data is in these panels.
> >
> > With relation "same", the limits are the same for all plots; with
> > relation "free" the limits depend on the data in the panel.  I thought
> > that relation "free" and specification of y-limits via prepanel
> > function (my previous trys) or as a list would do the trick:
> > scales=list(y=list(limits=list(panel-1-limits, panel-2-limits, ...))
> > but I cannot get it to work.
> >
> > I guess I am missing something basic ... any ideas or advice to give
> > up are greatly appreciated.
> >
> > Regards,
> > Boris.
> >
> >
> >
> > Hi Boris
> >
> > Not sure what you mean exactly
> > try
> >
> > library(latticeExtra)
> > useOuterStrips(dotplot(count ~ subject|risk*treatment,df))
> >
> > if you want to change the order of the subjects in each panel and an
> > index column and plot the index column instead of subject and change
> > the scales to suit.
> >
> > HTH
> >
> > Duncan
> >
> > Duncan Mackay
> > Department of Agronomy and Soil Science
> > University of New England
> > Armidale NSW 2351
> > Email: home: mackay at northnet.com.au At 07:54 16/02/2013, you wrote:
> >>Good afternoon, >>I would like to ask for help in controlling y-axis limits 
> >>and labels in >lattice doplots.  Unfortunately, the problem is somewhat 
> >>convoluted, >please bear with the long explanation. >>I would like to 
> >>create a 3x3 lattice of dotplots, say subject ~ count. >The plot is 
> >>conditioned on variables treatment and risk: subject ~ count >|treatment + 
> >>risk.  In the experiment,  not all subjects were exposed >to all 
> >>combinations of treatment and risk.  For each risk, I would like >to 
> >>show subject ~ count | treatment and order the subjects by the total 
> >>>count.  At the same time, I would like the x-axes to be the same in all 
> >>>panels and aligned by columns. >>Here is a sample data set: >># raw data 
> >>>df <- data.frame(subject=c('A','A','A','BB','BB','CCC','CCC','CCC',   
> >>>'DD','DD','A','A','A','FFFF','FFFF', >'A','A', 'B','B'), 
> >>>>risk=c('high','high','high','high',' high','high','high','high',
 >'med','med','med','med','med','med','med',
> >   >'low','low','low','low'), >>treatment=c('none 
> >','optX','optZ','none','optZ','none','optX','optZ',   
> >>>'none','optZ','none','optX','optZ','none','optZ',   
> >>'none','optX','none','optZ'), >count=c(5,10,2, 3,5,8,1,2, >3,7,10,2,5,15,2, 
> >>7,7,10,8)) ># re-level factors >df$risk <- 
> >factor(df$risk,levels=c('low','med','high')) >df $treatment <- 
> >factor(df$treatment,levels=c('none','optX','optZ'))   >>>## > df >##  
> >  subject risk treatment count >## 1        A high    
> >  none     5 >## 2        A high      optX  
> >  10 >## 3        A high      optZ     2 >## 
> >4       BB high      none     3 >## 5       
> >BB high      optZ     5 >## 6      CCC high    
> >  none     8 >## 7      CCC high      optX  
> >   1 >## 8      CCC high      optZ     2 >## 9 
> >     
 DD  med      none     3 >## 10      DD  med  
    optZ     7 >## 11       A  med      
none    10 >## 12       A  med      optX  Â
>   2 >## 13       A  med      optZ     5 >## 14 
>   FFFF  med      none    15 >## 15    FFFF  
>med      optZ
> >   2 >## 16       A  low      none     7 >## 
> >17       A  low      optX     7 >## 18    
> >   B  low      none    10 >## 19       B  
> >low      optZ     8 >>One way to plot the data is to break-up 
> >the data into sub-frames, one >frame for each risk, order subjects by total 
> >counts, create dotplots, >and merge with trellis.c().  This almost works 
> >but in the merged plot I >cannot decrease column spacing to be small 
> >enough.  Also, the output of >trellis.c() would not work with 
> >useOuterStrips() which I really like. >My code is in TRY ONE below. 
> >>>Another way to create the plot is specify y-limits for each panel with 
> >>custom prepanel and panel functions.  For each panel, the data-frame for 
> >>the panel row is isolated, subjects in the data-frame for the current >row 
> >are ordered by counts, panel y-limits are set to the re-ordered >levels, 
> >y-data for
 each panel is releveled, and data plotted with >standard panel.dotplot().  
This
>  somewhat works but lattice does not >honour
> >  the user-defined y-limits and labels are not correct.  I suspect 
> >>that it is not correct to use  y-relation="same" in this case but "free" 
> >>and "sliced" do not give correct results too. My code in in TRY TWO >below. 
> >>>If anybody can offer any assistance with this problem, it would be much 
> >>appreciated, >Sincerely, >Boris. >>>#### BEGIN TRY ONE - MERGE LATTICE 
> >PLOTS #### >library(lattice) >library(latticeExtra) >l ibrary(grid) >>for 
> >(irisk in levels(df$risk)) { ># subset data frame >df.irisk <- 
> >subset(df,risk==irisk) >># order subjects by total count; store levels of 
> >subjectx variables ># for later re-use in panel labels >df.irisk$subjectx <- 
> >df.irisk$subject[,drop=TRUE] >df.irisk$subjectx <- 
> >reorder(df.irisk$subjectx,df.irisk$count,sum) >a 
> >ssign(paste('sbjx.',irisk,sep=''),levels(df.irisk$subjectx))   >># create 
> >dotplot and store it in oltc.{irisk} variable >oltc.irisk <- 
> >dotplot(subjectx~count|treatment,data=df.irisk,
 >layout=c(3,1),type=c('p','h'),
> >   >xlim=c(-1,16),origin=0, >xlab="",ylab="") >as 
> >sign(paste('oltc.',irisk,sep=''),oltc.irisk) >}   >># combine everthing in 
> >one plot >oltc <- c(low=oltc.low,med=oltc.med,high=oltc.high) >pri nt(oltc) 
> >>># get rid of variable labels in middle and right column; decrease ># 
> >distance between columns.  But can't make inter-column spaces ># small 
> >enought and get rid of the panels in all but top rows. >laywid <- 
> >trellis.par.get('layout.widths') >laywid$between   <- -5 >laywid$axis.panel 
> ><- 0.7 >yscales <- list(labels=list(sbjx.low,NULL,NULL, >sbjx.med,N 
> >ULL,NULL, >sbjx.high,NULL,NULL)) >oltd <- 
> >update(oltc,scales=list(y=yscales), >par.setting 
> >s=list(layout.widths=laywid)) >print(oltd) >>### # END TRY ONE - MERGE 
> >LATTICE PLOTS #### >>#### BEGIN TRY TWO - CUSTOM PREPANEL AND PANEL 
> >FUNCTIONS #### >>prepanel.dotplot.x <- 
> >function(x,y,type,subscripts,...,data=NULL) { ># find data-frame that 
> >corresponds to the entire row of the plot >irisk <-
> >Â levels(data$risk[subscripts,drop=TRUE]) >idata <- 
> >subset(data,risk==irisk) >># in the sub-frame, order subjects by total 
> >counts >idata$subjectx <- 
> >reorder(idata$subject[,drop=TRUE],idata$count,sum)   >># set y-limits >ylim 
> ><- levels(idata$subjectx) >># increment packet counter and print new panel 
> >limits >pcknum = lattice.options()$packet.counter >lattice.option 
> >s(packet.counter=pcknum+1) >cat(paste(pcknum,":" ,paste(ylim,collapse=", 
> >")),sep="\n") >return(list(ylim=ylim)) >} >>pane l.dotplot.x <- 
> >function(x,y,type,subscripts,...,data=NULL) { ># get the sub-frame for the 
> >row >irisk <- levels(data$risk[subscripts,drop=TRUE]) >idata <- 
> >subset(data,risk==irisk) >># in the sub-frame, order subjects by total 
> >counts >idata$subjectx <- 
> >reorder(idata$subject[,drop=TRUE],idata$count,sum)   >># re-order 
> >y-variable to "correct levels" >y <- 
> >factor(as.character(y),levels(idata$subjectx)) >  ># print levels of the 
> >releveled subjects - should be the same ># as
 the output of
> >  prepanel.dotplot.x >pnlnum <- panel.number() 
> >>cat(paste(pnlnum,':',paste(level s(y),collapse=", ")),sep="\n") >># call 
> >standard dotplot >panel.dotplot(x,y,...) >} >># the data is plotted 
> >correctly but the labels and limits are not >correct 
> >>lattice.options(packet.counter=1)   >oltc <- 
> >dotplot(subject~count|treatment+risk,data=df, >l 
> >ayout=c(3,3),subscripts=TRUE, >prepanel=function (...) 
> >>{prepanel.dotplot.x(...,data=df)}, >panel 
> >=function(...){panel.dotplot.x(...,data=df)}, >s 
> >cales=list(y=list(relation='same')), >drop.unuse d.levels=FALSE, 
> >>xlim=c(0,16)) >print(oltc) >>## ## END TRY TWO - CUSTOM PANEL AND PREPANEL 
> >FUNCTIONS #### >>_________________________________________ _____ >R-help at 
> >r-project.org mailing list >https://stat.ethz.ch/mailman/listinfo/r-he lp 
> >>PLEASE do read the posting guide 
> >http://www.R-project.org/posting-guide.html >and   provide commented, 
> >minimal, self-contained, reproducible code.
> >         [[alternative HTML version deleted]]
> >
> > ______________________________________________
> > R-help@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
>         [[alternative HTML version deleted]]
> 
> 
> ______________________________________________
> R-help@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.
        [[alternative HTML version deleted]]


______________________________________________
R-help@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.