print.TwoTrellisColumns: Print two conformable trellis plots in adjacent columns with user control of widths.

Description

Print two conformable trellis plots in adjacent columns with user control of widths. Left y tick-labels and left.strip are removed from the right-hand plot.

Usage

as.TwoTrellisColumns5(left,  ## left  is the left trellis object
                      right, ## right is the right trellis object
                      ## Both left and right must have identical
                      ## settings for number and size of vertical panels,
                      ## left-axis labels, number of lines in main, sub, legend.
                      ...,
                      pw=c(.3, .30, .01, .30, .09),
                      px=list(
                        LL=c(0, pwc[1]),
                        LP=pwc[1:2],
                        ML=pwc[2:3],
                        RP=pwc[3:4],
                        RL=pwc[4:5]),
                      pwc=cumsum(pw),
                      strip.left=TRUE,
                      y.tck=c(0,0)
                      )
# S3 method for TwoTrellisColumns5
print(x, px=attr(x, "px"), ...)
leftLabels.trellis(x)
rightLabels.trellis(x)
panelOnly.trellis(x, strip.left=FALSE, y.tck=0)
mainSubLegend.trellis(x)
emptyRightAxis(x)

Arguments

left, right

Conformable "trellis" objects. Both must have the identical settings for number and size of vertical panels, left-axis labels, number of lines in main, sub, legend.

"trellis" object.

These are used x-values used in the position argument of the print.trellis function. The default (constructed from the pw argument) makes the Left and Right panels the same width and the Middle containing the y-axis is given the remainder. Overlapping is permitted. The appearance depends on the width of the graphics device.

pw, pwc

pw vector of five positive numbers that sum to 1. These are the relative widths of the five sections of the result: LeftLabels, LeftPanel, MainSubLegend, RightPanel, RightLabels. pwc is the cumulative sum of pw. pwc is expanded in the px argument to the x values used in the position argument of the print.trellis function.

strip.left

See barchart.

y.tck

A vector of one or two numeric values. This will be used as the y tck value for the right column of panels. See 'tck' in barchart for details.

…

Other arguments are ignored.

Value

A "TwoTrellisColumns5" object, consisting of a list containing the constructed left, middle, and right trellis objects, and an attribute containing the px value.

Details

as.TwoTrellisColumns5 constructs a "TwoTrellisColumns5" object, which is a list of five trellis objects named "LL", "LP", "ML", "RP", "RL". LL is the left labels from the left input object. LP is the panels from the left input object. ML is the middle labels from the left object; these are the main title, sub title, and legend. RP is the panels from the right input object. RL is the right labels from the right input object.

print.TwoTrellisColumns5 is a print method for a "TwoTrellisColumns5" object. It takes left-to-right positioning information from the "px" attribute of its argument x or from an input argument. The numbers are used as the "x" information for the position argument to the print.trellis method.

emptyLeftAxis, leftLabels.trellis, rightLabels.trellis, panelOnly.trellis, mainSubLegend.trellis, emptyLeftStrip, emptyRightAxis are functions which blank out the various components of the trellis argument and retains their vertical spacing.

Examples

Run this code

# NOT RUN {

## These are based on the Professional Challenges example in ?likert
data(ProfChal)
levels(ProfChal$Subtable)[6] <- "Prof Recog" ## reduce length of label

## initial ordering of Question factor
PCC <- likert(Question ~ . | Subtable, ProfChal, ylab=NULL,
              rightAxis=TRUE,
              layout=c(1,6),
              strip=FALSE,
              strip.left=strip.custom(bg="gray97"),
              par.strip.text=list(cex=.7),
              scales=list(y=list(relation="free")),
              main="Is your job professionally challenging?")

## initial ordering of Question factor
PCP <- likert(Question ~ . | Subtable, ProfChal, ylab=NULL,
              as.percent=TRUE,
              layout=c(1,6),
              strip=FALSE,
              strip.left=strip.custom(bg="gray97"),
              par.strip.text=list(cex=.7),
              scales=list(y=list(relation="free")),
              main="Is your job professionally challenging?")

# }
# NOT RUN {
## default equal widths of the two panels
as.TwoTrellisColumns5(PCP, PCC)  ## 11in x 7in

## make left panel twice as wide as right panel
as.TwoTrellisColumns5(PCP, PCC, pw=c(.3, .4, .01, .2, .09))  ## 11in x 7in
##                                   --------------------    ## sum to 1.00

## make left panel twice as wide as right panel, and control position of main and legend
as.TwoTrellisColumns5(PCP, PCC,  ## 11in x 7in
                      px=list(
                        LL=c(.00,  .30),
                        LP=c(.30,  .70),
                        ML=c(.60,  .61),  ## arbitrary,
                                          ## visually center the labels and legend
                        RP=c(.71,  .91),
                        RL=c(.91, 1.00)))
# }
# NOT RUN {
## Size that works in default 7x7 window.  7x7 is not recommended for
## this example because most of the space is used for labeling and not
## much for the panels containing the data.  Use the px values for the
## 11x7 illustrated above in the dontrun section.

as.TwoTrellisColumns5(PCP, PCC,  ## 7in x 7in
                      px=list(
                        LL=c(.00,  .50),
                        LP=c(.50,  .70),
                        ML=c(.50,  .51),  ## arbitrary,
                                          ## visually center the labels and legend
                        RP=c(.71,  .87),
                        RL=c(.87, 1.00)))

## Ordering the rows by the lengths of the positive bars and also
## put percents and counts on the same plot.
## The easiest way is to use the LikertPercentCountColumns function:

LikertPercentCountColumns(Question ~ . | Subtable, ProfChal,
                          layout=c(1,6), scales=list(y=list(relation="free")),
                          ylab=NULL, between=list(y=0),
                          strip.left=strip.custom(bg="gray97"), strip=FALSE,
                          par.strip.text=list(cex=.7),
                          positive.order=TRUE,
                          main="Is your job professionally challenging?")

# }
# NOT RUN {
## Ordering the rows by the lengths of the positive bars and also
## putting percents and counts on the same plot requires coordination.
## The easiest way is to order the original tables of counts by the
## order of the percent plot.

percentPlot <- likert(Question ~ . | Subtable, ProfChal,
                      as.percent=TRUE,
                      layout=c(1,6), scales=list(y=list(relation="free")),
                      ylab=NULL, between=list(y=0),
                      strip.left=strip.custom(bg="gray97"), strip=FALSE,
                      par.strip.text=list(cex=.7),
                      positive.order=TRUE,
                      main="Is your job professionally challenging?")
## percentPlot
pct.order <- percentPlot$y.limits[[1]]

ProfChal2 <- ProfChal
ProfChal2$Question <- factor(ProfChal2$Question, levels=rev(pct.order))

countPlot   <- likert(Question ~ . | Subtable, ProfChal2,
                      layout=c(1,6),
                      rightAxis=TRUE,
                      scales=list(y=list(relation="free"),
                                  x=list(at=c(0, 250, 500))),
                      ylab=NULL, between=list(y=0),
                      strip.left=strip.custom(bg="gray97"), strip=FALSE,
                      par.strip.text=list(cex=.7),
                      main="Is your job professionally challenging?")
## countPlot
levels(ProfChal$Subtable)[6] <-
      "Attitude\ntoward\nProfessional\nRecognition" ## Restore original label

## Size that works in default 7x7 window.  7x7 is not recommended for
## this example because most of the space is used for labeling and not
## much for the panels containing the data.  Use the px values for the
## 11x7 illustrated above in the dontrun section.

as.TwoTrellisColumns5(percentPlot, countPlot,  ## 7in x 7in
                      px=list(
                        LL=c(.00,  .50),
                        LP=c(.50,  .70),
                        ML=c(.50,  .51),  ## arbitrary,
                                          ## visually center the labels and legend
                        RP=c(.71,  .87),
                        RL=c(.87, 1.00)))
# }