require(alr3)
data(heights)
The heights dataset contains 1375 observations of heights of a mother and her daughter. This dataset was compiled by the statistician Karl Pearson in the late 1800s. To make this a more interesting example of a permutation test, we will limit our sample size and only consider a random sample of 200 rows of the dataset. We will start by describing the data visually with a simple graphic.
summary(heights)
## Mheight Dheight
## Min. :55.4 Min. :55.1
## 1st Qu.:60.8 1st Qu.:62.0
## Median :62.4 Median :63.6
## Mean :62.5 Mean :63.8
## 3rd Qu.:63.9 3rd Qu.:65.6
## Max. :70.8 Max. :73.1
Now let's plot the data using ggplot2. First we're going to plot side by side boxplots of the mother's heights next to the daughter's heights.
require(ggplot2)
## Loading required package: ggplot2
heightStack <- data.frame(height = c(heights$Mheight, heights$Dheight), who = factor(c(rep("Mother",
1375), rep("Daughter", 1375)), levels = c("Mother", "Daughter")))
ggplot(heightStack, aes(who, height)) + geom_boxplot(outlier.colour = "darkgreen",
aes(fill = who))
Now we'll do a scatter plot of mother's heights versus daughter's heights. The first line in the code below changes the theme. qplot acts similarly to base R command plot.
theme_set(theme_bw()) ## sets default white background theme for ggplot
qplot(Mheight, Dheight, data = heights, col = "red", alpha = 0.5) + theme(legend.position = "none")
We can also do this using ggplot.
theme_set(theme_bw()) ## sets default white background theme for ggplot
ggplot(heights, aes(Mheight, Dheight)) + geom_point(col = "red", alpha = 0.5) +
stat_smooth()
