Topic 6 Graphing and data visualisation with R

6.1 Presenting data visually

6.2 Using GGplot to make graphs

6.3 By the end of this section, you will be able to:

  • Describe the ggplot “grammar of visualisation”: coordinates and geoms
  • Write a graph function to display multiple variables on a plot
  • Amend the titles and legends of a plot
  • Save plots in PDF or image formats

6.4 The “grammar of visualisation”

  • Graphs are made up of 3 components:
    • A dataset
    • A coordinate system
    • Visual marks to represent data (geoms)

The “grammar of visualisation” #2

  • In the above example, the dataset is the studentData that we used previously.
  • The grades variable is mapped to the X axis
  • The hoursOfStudy variable is mapped to the Y axis

6.5 How to code a graph

  • The graph is created using the following code:

  • In this code, we specify the dataset, the variables for the X and Y axes and the geom that will represent the data points visually (in this case, each datum is a point)

6.6 The graph output

library(ggplot2)

ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) + geom_point()

6.7 Changing the geoms leads to different visualisations

  • If we change from points to lines, for example we get a different plot:
library(ggplot2)

ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) + geom_line()

6.8 It is possible to represent more variables on the plot

  • By specifying that colours of our points should be attached to the route variable, the data is now colour-coded
library(ggplot2)

ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) + geom_point(aes(color = route))

6.9 It is possible to represent more variables on the plot #2

  • By specifying that size of our points should be attached to the satisfactionLevel variable, the size of the points adjusts
library(ggplot2)

ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) + geom_point(aes(color = route, size=satisfactionLevel))
## Warning: Using size for a discrete variable is
## not advised.

6.10 It is possible to represent more variables on the plot #3

  • By specifying that shape of our points should be attached to the hasDependents variable, the shape of the points changes accordingly
library(ggplot2)

ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) + geom_point(aes(color = route, size=satisfactionLevel, shape=hasDepdendants))
## Warning: Using size for a discrete variable is
## not advised.

6.11 Plotting summaries of data

  • We can summarise the data (e.g. get the mean or sd) using the stat_summary() function
  • Below we are making a bar chart with the mean grade for each route
ggplot(data=studentData, aes(x=route, y= grades, fill=route)) + stat_summary(fun.y = "mean", geom = "bar")
## Warning: The `fun.y` argument of
## `stat_summary()` is deprecated as of
## ggplot2 3.3.0.
## ℹ Please use the `fun` argument
##   instead.

6.12 Changing the axis labels and title on a plot

We can change the axis labels and title using the labs() command:

labs(x="Student Grade", y="Hours of Study", title = "Scattterplot of student data")
library(ggplot2)

ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) + geom_point(aes(color = route, size=satisfactionLevel, shape=hasDepdendants)) + labs(x="Student Grade", y="Hours of Study", title = "Scattterplot of studentdata")
## Warning: Using size for a discrete variable is
## not advised.

6.13 Changing the legend on a plot

To change the legend, we use the labs() command too, and reference the relevant property (e.g. size, shape, colour)

labs(x="Student Grade", y="Hours of Study", title = "Scattterplot of student data", color="Route of study", size="Satisfaction level", shape="Has dependents?")
library(ggplot2)

  ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) + 
    geom_point(aes(color = route, size=satisfactionLevel, shape=hasDepdendants)) +
  labs(x="Student Grade", y="Hours of Study", title = "Scattterplot of studentdata", color="Route of study", size="Satisfaction level", shape="Has dependents?")
## Warning: Using size for a discrete variable is
## not advised.

6.14 Storing plots to be recalled later

  • Plots can be assigned to objects in R and recalled later, just like any other piece of data
library(ggplot2)

## Create plot and store it as "myPlot" object

myPlot <- ggplot(data=studentData, aes(x=grades,y=hoursOfStudy)) +
  geom_point(aes(color = route, size=satisfactionLevel, shape=hasDepdendants)) +
  labs(x="Student Grade", y="Hours of Study", title = "Scattterplot of studentdata", color="Route of study", size="Satisfaction level", shape="Has dependents?")

6.15 Recalling a stored plot

 #Recall myPlot
myPlot
## Warning: Using size for a discrete variable is
## not advised.

6.16 Saving plots # 1

  • Plots can be save using the export button in the plots tab

6.17 Plots can also be saved using code

  • You might want to include code to save your plot in a script, for example
  • This can allow greater control over the output file and plot dimensions:
ggsave(plot= myPlot, file="myPlot.pdf", width = 4, height = 4)
## Warning: Using size for a discrete variable is
## not advised.
ggsave(plot= myPlot, file="myPlot.png", width = 4, height = 4, units="cm", dpi=320)
## Warning: Using size for a discrete variable is
## not advised.