Session Objectives

Data Visualisation with ggplot2

• Introduction to ggplot2
• Boxplot
• Grouped Boxplot
• Scatterplot
• Means plot

Setup

Tasks:

1. Open the slides for today

2. Open/create your seminRs project

3. Download the Rmd document to the r_docs folder of your seminRs project & open it

You can download the Solutions Rmd to check your answers later

Intro to ggplot2

Figures/graphs/plots are a useful way to look at your data

• Quickly see general trends of data with minimal effort

• Easy check of data

• Gives readers a general overview of your results/data

ggplot2 is a really useful package for data visualisation

• Part of tidyverse

• Follows The Grammar of Graphics

• Plots are built in layers from 3 key components: data, coordinate system, & geometric elements (geoms)

• Super customisable

Using ggplot()

• Built layer by layer +

• Variables are mapped onto elements of the plot as 'aesthetics'
  • Use aes() to include variable(s) as an aesthetic

• Geoms are 'visual marks' that represent data points e.g.

  • geom_bar() – creates a layer of bars
  • geom_point() – creates a layer of data points
  • geom_histogram() – creates a layer with a histogram
  • geom_text() – creates layer with text on

& many more options...

General Process

1. Create base layer by specifying the data & the variables to map onto the plot

2. Add geom layer (to display our data points)

3. Add/edit visual properties (scale, colours, shapes etc.)

4. Add semantic properties (e.g. title, labels)

5. Add a theme to make our plot pretty

Today we're going to follow this process, to build different types of common graphs!

Basic Boxplot

peng_boxplot <- ggplot2::ggplot(peng_data, aes(island, bill_length_mm))
peng_boxplot +
  geom_boxplot() +
  labs(title = "Penguin Bill Length Split by Island", x = "Island", y = "Bill Length (mm)") + 
  theme_classic()Copy Code

Grouped Boxplot

peng_boxplot_2 <- ggplot2::ggplot(peng_data, aes(island, bill_length_mm, fill = sex))
peng_boxplot_2 +
  geom_boxplot() +  
  labs(title = "Penguin Bill Length Split by Island and Sex", x = "Island", y = "Bill Length (mm)", fill = "Sex") + 
  theme_minimal()Copy Code

Scatterplots

peng_scatter <- ggplot2::ggplot(peng_data, aes(bill_length_mm, bill_depth_mm))
peng_scatter +
  geom_point() +
  labs(title = "Scatterplot of Bill Length & Bill Depth", x = "Bill Length (mm)", y = "Bill Depth (mm)") + 
  theme_bw()Copy Code

peng_scatter_2 <- ggplot2::ggplot(peng_data, aes(bill_length_mm, bill_depth_mm))
peng_scatter_2 +
  geom_point(colour = "dark blue", size = 3, shape = 20, alpha = .5) +
  labs(title = "Scatterplot of Bill Length & Bill Depth", x = "Bill Length (mm)", y = "Bill Depth (mm)") + 
  theme_bw()Copy Code

Means Plots

peng_means <- ggplot2::ggplot(peng_data, aes(species, flipper_length_mm))
peng_means +
  stat_summary(fun = "mean", geom = "point", size = 4, colour = "pink") +  
  labs(title = "Mean Flipper Length per Penguin Species", x = "Species", y = "Flipper Length (mm)") + 
  theme_minimal()Copy Code

Don't Be Misleading!

Changing the Coordinate System

peng_means <- ggplot2::ggplot(peng_data, aes(species, flipper_length_mm))
peng_means +
  stat_summary(fun = "mean", geom = "point", size = 4, colour = "pink") +  
  labs(title = "Mean Flipper Length per Penguin Species", x = "Species", y = "Flipper Length (mm)") + 
  coord_cartesian(ylim = c(0, 250)) +
scale_y_continuous(breaks = seq(0, 250, 50)) +
  theme_minimal()Copy Code

General Tips

Think about what makes a good graph when creating your own:

• Clear & tidy
• Simple – presents a lot with a little!
• Coherent
• Accurate – doesn’t distort results
• Informative captions/titles & labels
• Is referred to in the text ("see Figure 1")

To Avoid Problems

• Run & inspect every change you make
• Be aware that creating plots with ggplot is often a long process
• A really common error is forgetting the + for each layer
• Set code chunk options to echo = FALSE (plot is displayed but the code isn't)