# simple example
rmb(formula = Freq ~ Type + Infl + Cont + Sat, data = housing, gap.mult = 2,
col.vars = c(FALSE, TRUE, TRUE, FALSE), label.opt = list(abbrev = 3))
# with sqrt-transformation and horizontal splits only
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, gap.mult = 2, col.vars = c(TRUE,
TRUE, TRUE, TRUE), freq.trans = "sqrt", label.opt = list(abbrev = 3))
# a generalized spineplot with the first category highlighted
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, spine = TRUE, cat.ord = 1,
mult = 2, col.vars = c(1, 3, 4), freq.trans = list("sqrt", 3), label.opt = list(abbrev = 2))
# a generalized spineplot with all categories highlighted in a changed order
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, spine = TRUE, cat.ord = c(3,
1, 2), gap.mult = 2, col.vars = c(TRUE, FALSE, TRUE, TRUE), freq.trans = "sqrt",
label.opt = list(abbrev = 3))
# the barchart version only for categories 1 and 3
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, cat.ord = c(1, 3),
gap.mult = 2, col.vars = c(TRUE, FALSE, TRUE, TRUE), freq.trans = "sqrt",
label.opt = list(abbrev = c(4, 1, 1, 1)))
# with equal widths
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, eqwidth = TRUE, gap.mult = 2,
col.vars = c(TRUE, FALSE, TRUE, TRUE), label.opt = list(abbrev = 2, lab.tv = TRUE))
“# —– models and residuals —– # # using the logistic model: Sat by Type only
# residual shadings and expected values ```
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, gap.mult = 2, col.vars = c(TRUE,
FALSE, TRUE, TRUE), label.opt = list(abbrev = 3), expected = list(c(1, 2,
3), c(1, 4)), model.opt = list(use.expected.values = TRUE, resid.display = "color"))
# residual values without shadings
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, gap.mult = 2, col.vars = c(TRUE,
FALSE, TRUE, TRUE), label.opt = list(abbrev = 3), expected = list(c(1, 2,
3), c(1, 4)), model.opt = list(resid.display = "values"))
# residual shadings and expected values
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, gap.mult = 2, col.vars = c(TRUE,
FALSE, TRUE, TRUE), label.opt = list(abbrev = 3), expected = list(c(1, 2,
3), c(1, 4)), model.opt = list(use.expected.values = TRUE, resid.display = "color"))
# barcharts with residual shadings and values
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, gap.mult = 2, col.vars = c(TRUE,
FALSE, TRUE, TRUE), label.opt = list(abbrev = 3), expected = list(c(1, 2,
3), c(1, 4)))
# spineplots with residual shadings and values
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, spine = TRUE, gap.mult = 2,
col.vars = c(TRUE, FALSE, TRUE, TRUE), label.opt = list(abbrev = 3), expected = list(c(1,
2, 3), c(1, 4)))
# piecharts with residual shadings and values
rmb(formula = ~Type + Infl + Cont + Sat, data = housing, circular = TRUE, gap.mult = 2,
col.vars = c(TRUE, FALSE, TRUE, TRUE), label.opt = list(abbrev = 3), expected = list(c(1,
2, 3), c(1, 4)))
# ----- using an ftable to create the plot ----- #
tt = xtabs(Freq ~ Type + Cont + Infl + Sat, data = housing)
ft = ftable(tt, col.vars = c(1, 4))
rmb(tt, col = "seq")
rmb(ft, col = "seq", col.opt = list(h = 140))
# ----- using a glm model ----- #
mod1 <- glm(Freq ~ Type * Infl * Cont + Type * Sat, data = housing, family = poisson)
rmb(mod1, circular = TRUE, gap.mult = 2, col.vars = c(TRUE, FALSE, TRUE, TRUE),
label.opt = list(abbrev = 3), model.opt = list(use.expected.values = TRUE))
# ----- the numeric mode and cuts ----- #
# only three cuts to show how it works
data(olives)
rmb(~palmitoleic + stearic + Region, data = olives, cut = c(3, 3, 0))
# a minority of extreme observations mess the display up:
require(ggplot2)
data(diamonds)
diamonds$lprice <- log(diamonds$price)
rmb(~depth + table + lprice, data = diamonds, eqwidth = TRUE, spine = TRUE,
cut = c(36, 36, 5), col = "seq", num.mode = TRUE)
# we can zoom in via innerval:
rmb(~depth + table + lprice, data = diamonds, circular = TRUE, cut = c(36, 36,
5), col = "div", innerval = 0.95, num.mode = TRUE, freq.trans = "log")