compactr (version 0.1)

eplot: Create an empty plot with compact axis notation

Description

The eplot() function draws an empty plot to which the user can add points, lines, text, etc. The axis notation is more compact than the defaults for the plot() function. Also, axis and label are appropriately suppressed when the plot occurs as part of a matrix. The aplot() function simply calls eplot() again, using the same arguments (with the exception of main).

Usage

eplot(xlim, ylim, xlab = NULL, ylab = NULL, main = NULL, text.size = 1, tick.length = 0.02, xpos = -0.7, ypos = -0.5, xat = NULL, yat = NULL, xticklab = NULL, yticklab = NULL, xlabpos = 1.5, ylabpos = 1.5, annx = TRUE, anny = TRUE, box = TRUE)

Arguments

xlim

the x limits (x1, x2) of the plot.

ylim

the y limits of the plot.

xlab

a label for the x axis, defaults to empty space.

ylab

a label for the y axis, defaults to empty space.

main

a label for the subplot. Intended for labeling a each plot in a matrix. If you need a title for the entire matrix of plots, or a single plot, I recommend using a call to the mtext() function.

text.size

a numerical value giving the amount by which axis notation should be magnified. Reasonable values range from about 0.5 to 2.

tick.length

the length of tick marks as a fraction of the smaller of the width or height of the plotting region. Reasonable values range from about 0.01 to 0.1.

xpos,ypos

controls the distance from the tick labels to the axis. Reasonable values range from about -1 to 1.

xat,yat

the location of the tick marks along the axes. If "none," then the axis will not be annotated.

xticklab,yticklab

the labels for the tick marks. A character vector the length of xat and yat.

xlabpos,ylabpos

controls the distance from the axis labels to the axes. Reasonable values range from about 1 to 3.

annx,anny

include annotations for x and y axes?

box

should a box be plotted?

Details

This function simply draws an empty plot with compact axis notation, to which the user can add points, lines, text, and so on. Also, if the plot appears as part of a matrix, the x-axis notation is suppressed unless the plot appears along the bottom row and the y-axis notation is suppress unless the plot appears along the left column.

Examples

### Plot 0: illustrating the purpose

# run these lines one at a time to see what happens
par(mfrow = c(2,2))
eplot(xlim = c(-1, 1), ylim = c(0, 10))
aplot(main = "Hey Look! No axis labels.")
aplot(main = "But this one has them?!")
aplot(main = "And this one does just what you'd expect!")
# after a call to eplot() or aplot(), I just add
# whatever I want to the plot.


### Plot 1: a simple scatter plot

set.seed(1234)
x <- rnorm(100)
y <- x + rnorm(100)

par(mfrow = c(1,1), mar = c(3,3,1,1), oma = c(0,0,2,0))
eplot(xlim = c(min(x), max(x)), ylim = c(min(y), max(y)),
           xlab = "Explanatory Variable", ylab = "Outcome Variable")
points(x, y)
abline(lm(y ~ x), lwd = 3, col = "red")
mtext("A Clever Title", outer = TRUE)


### Plot 2: a matrix of scatter plots

# simulation multilevel data
set.seed(1234)
group <- rep(1:11, each = 15)
a <- rnorm(length(unique(group)), sd = 1)
b <- rnorm(length(unique(group)), mean = 1, sd = .3)
x <- rnorm(length(group))
y <- a[group] + x*b[group] + rnorm(length(group))

## estimate random effects models and pull out the estimates
#library(lme4)
#hier <- lmer(y ~ x + (1 + x | group))
#a.hat <- fixef(hier)[1] + ranef(hier)$group[, 1]
#b.hat <- fixef(hier)[2] + ranef(hier)$group[, 2]

# draw plot
par(mfrow = c(3,4), mar = c(.75,.75,.75,.75), oma = c(4,4,4,1))
for (i in 1:11) {
  eplot(xlim = c(min(x), max(x)), ylim = c(min(y), max(y)),
             xlab = "Explanatory Variable", ylab = "Outcome Variable",
             main = paste("Group", i))
  points(x[group == i], y[group == i])
  #abline(a = a.hat[i], b = b.hat[i])
  abline(lm(y[group == i] ~ x[group == i]), lty = 3)
}

# add an overall title
mtext("Comparing Partial Pooling and No Pooling", outer = TRUE, line = 2)


### Plot 3: a matrix of scatter plots using aplot() and addxaxis()

# use the same estimates as before

# draw the first plot with eplot()
par(mfrow = c(3,4), mar = c(.75,.75,.75,.75), oma = c(4,4,4,1))
eplot(xlim = c(min(x), max(x)), ylim = c(min(y), max(y)),
           xlab = "Explanatory Variable", ylab = "Outcome Variable",
           main = "Group 1")

# then add stuff
points(x[group == 1], y[group == 1])
#abline(a = a.hat[1], b = b.hat[1])
abline(lm(y[group == 1] ~ x[group == 1]), lty = 3)
legend(par("usr")[1], par("usr")[4],
       legend = c("partial pooling", "no pooling"), lty = c(1, 3),
       bty = "n", bg = NA, cex = .8)

# draw the rest with aplot()
for (i in 2:11) {
  aplot(main = paste("Group", i))
  # since we don't plan to have bottom right plot,
  # let's add an axis to the one above
  if (i == 9) { addxaxis() }
  points(x[group == i], y[group == i])
  #abline(a = a.hat[i], b = b.hat[i])
  abline(lm(y[group == i] ~ x[group == i]), lty = 3)
}
mtext("Comparing Partial Pooling and No Pooling", outer = TRUE, line = 2)