In jr-packages/jrSouth: Advanced R graphics

Getting Started

First load the ggplot2 package

library("ggplot2")

and the OK Cupid data set

data(okcupid, package = "jrSouth")

Scatter plots

1. Let's start by creating a basic scatter plot of the heights and ages. We do this using the geom_point() command

# alpha makes the points transparent
ggplot(data = okcupid) + 
  geom_point(aes(x = age, y = height), alpha = 0.2)

1. To save typing, we're going to store the original ggplot object as a variable

g = ggplot(data = okcupid)
g1 = g + geom_point(aes(x = age, y = height), alpha = 0.2)

So now running g1 will produce the graph

g1

1. The arguments x and y here are called aesthetics. What do you think happens if you omit the y aesthetic?

# this gives an error

1. For geom_point(), both the x and y aesthetics are required. This particular geom has other aesthetics: shape, colour, size and alpha.^[These are available for most geoms. For a collection of aesthetics see the relevant help pages.] For instance we can specify that we wish to map the variable sex to a colour aesthetic by including it inside aes()

   r g + geom_point(aes(x = age, y = height, colour = sex))

2. Change colour = sex to colour = height. Why do you think there's a change in the legend?

Bar plots

geom_bar() can be used to create a bar chart. It requires only one aesthetic and that is x. For the provided aesthetic, the frequencies will be calculated and shown as a bar. For example

g + geom_bar(aes(x = body_type))

1. Change the axis labels to "Body Type" and "Total" by adding two more layers, using + xlab("Body Type") and + ylab("Total")

g + geom_bar(aes(x = body_type)) + 
  xlab("Body Type") + 
  ylab("Total")

1. Split the graph up into the two genders (hint: use colour and fill).

# What happens if you only have colour or only fill?
g + geom_bar(aes(x = body_type, colour = sex, fill = sex))

1. With such long labels, it might make more sense to rotate the graph such that the bars and labels are horizontal. Switch the x and y axes using a coord_flip() layer added to your graph.

g + geom_bar(aes(x = body_type, colour = sex, fill = sex)) + 
  coord_flip()

1. I am not too keen on how the female and male bars are displayed on top of each other. The argument to change these is position. The default is stack, for example we can put the bars next to each other using

   r g + geom_bar(aes(x = body_type, colour = sex, fill = sex), position = "dodge") + coord_flip()

   Other values you might try here are position = fill, position = jitter, or position = identity.

2. What does the fill position argument do?

g + geom_bar(aes(x = body_type, colour = sex, fill = sex), position = "fill") + 
  coord_flip()
# puts the values on a common scale (all sum to 1)

Box plots

1. Box plots are a great way to visualise the shape of a distribution of some variable. Start by creating a boxplot of peoples ages in the okcupid data. The x = 1 in the code below lets us have just a single boxplot for all ages

g = ggplot(okcupid) 
g + geom_boxplot(aes(x = 1, y = age))

1. Switch out the x = 1 for x = smokes to get a boxplot for each group

g + geom_boxplot(aes(x = smokes, y = age))

Copy cat

data(mpg, package="ggplot2")
# mpg$drv = as.character(mpg$drv)
# mpg[mpg$drv == "f",]$drv = "Front"
# mpg[mpg$drv == "r",]$drv = "Rear"
# mpg[mpg$drv == "4",]$drv = "4wd"
# mpg$drv = factor(mpg$drv,
#    levels = c("Front", "Rear", "4wd"))

g = ggplot(data=mpg, aes(x=displ, y=hwy))
g1 = g + geom_point() + stat_smooth(linetype=2) +
  xlab("Displacement") + ylab("Highway mpg")
g2 = g + geom_point() + stat_smooth(aes(colour=drv))

g1

The aim of this section is to recreate the graphics in figure 1 and 2. Feel free to experiment. To begin, load the package

library("ggplot2")

\noindent and the mpg data set

data(mpg, package="ggplot2")
dim(mpg)

1) Figure 1: Create a scatter plot of engine displacement, displ, against highway mpg, hwy. To get started:

ggplot(data=mpg, aes(x=displ, y=hwy)) +
  geom_point() + xlab("Displacement")

Now add a dashed loess line and change the $y$-axis label. Hint: try stat_smooth() and ylab('New label').

g2

g1 = g + geom_point() + stat_smooth(linetype=2) +
  xlab("Displacement") + ylab("Highway mpg")

2) Figure 2: Using stat_smooth(), add a loess line conditional on the drive.

g2 = g + geom_point() + stat_smooth(aes(colour=drv))

jr-packages/jrSouth documentation built on May 23, 2019, 8:50 a.m.