qmosaic: Mosaic plot. Create a mosaicplot using a formula (as...
In ggobi/cranvas: Interactive statistical graphics based on Qt

Description Usage Arguments Author(s) Examples

Interactive elements for mosaic plots are arrow keys for navigating through the mosaic hierarchy: arrow up reduces complexity of the mosaic by one variable, arrow down increases the complexity by one, if possible. Arrow left and right rotate a previously included variable into the last split position. Conditioning/Unconditioning is done with keys 'C' and 'U' Keys 'B' and 'S" switch to bar and spine representation, respectively Key 'R' rotates the last split variable between horizontal and vertical display.

1 2	qmosaic(data, formula, divider = productplots::mosaic(), cascade = 0, scale_max = TRUE, na.rm = FALSE, subset = NULL, colour = "grey30", main = NULL, ...)

`data`	a mutaframe which is typically built upon a data frame along with several row attributes
`formula`	a formula to describe order in which variables go into the mosaicplot. The first variables are the ones visually most important, i.e. Y ~ X1 + X2 + X3 first splits according to X3, then according to X2, then according to X1
`divider`	structure of the split in each direction. Choices are "hbar", "vbar" for horizontal/vertical barcharts, "hspine", "vspine" for horizontal/vertical spineplots.
`cascade`	parameter for prodplot in package productplots
`scale_max`	parameter for prodplot in package productplots
`na.rm`	handling of missing values, defaults to FALSE
`subset`	parameter for prodplot -
`colour`	fill colour of rectangles - only used if colour is not used in the data
`main`	parameter for prodplot
`...`

Heike Hofmann

library(cranvas)
library(productplots)
library(plyr)


## Simpson on board the Titanic?!
data(Titanic)
titanic <- as.data.frame(Titanic)
titanic$Passenger <- titanic$Class != "Crew"
qtitanic <- qdata(titanic, color = Survived)

qtitanic$.brushed <- qtitanic$Survived == "Yes"
## passengers have a higher probability of survival than crew members ...
qmosaic(qtitanic, Freq ~ Passenger, divider = c("hspine"))
## unless sex is taken into account: the conditional association between Passenger
## and Survival changes to the opposite: both male and female crew members had
## better changes of survival than their counterparts on the passenger side.
qmosaic(qtitanic, Freq ~ Passenger | Sex, divider = c("hspine", "hspine"))

## this is an example of the Simpson's paradox, and is due to the strong
## association between crew membership and gender.
qmosaic(qtitanic, Freq ~ Sex + Passenger, divider = c("vspine", "hspine"))

####################### 
data(happy)
qhappy <- qdata(happy, color = happy)

qmosaic(qhappy, ~sex + marital, divider = c("vspine", "hspine"))
qmosaic(qhappy, ~marital + happy + year, divider = c("vspine", "hspine", "hspine"))
qmosaic(qhappy, ~marital + degree, divider = c("vspine", "hspine"))
qmosaic(qhappy, ~happy, divider = c("hbar"))

qbar(sex, qhappy)


qmosaic(qtitanic, Freq ~ Sex + Age + Class)
qbar(Survived, data = qtitanic)

qmosaic(qtitanic, Freq ~ Survived, "hbar")
qmosaic(qtitanic, Freq ~ Age, "hbar")
qmosaic(qtitanic, Freq ~ Sex, "hbar")
qmosaic(qtitanic, Freq ~ Class, "hbar")

qmosaic(qtitanic, Freq ~ Sex + Survived | Class, divider = c("vspine", "hspine", 
    "hspine"))
data(happy)
qhappy <- qdata(happy)
qmosaic(qhappy, ~happy, "hbar")
qmosaic(qhappy, ~happy + age, c("hspine", "vspine"))

cranvas_off()