Lab 5

Overview

This lab focuses on refining visualizations for comparing categories through effective ordering, theme customization, and strategic annotations. You’ll practice creating publication-ready visualizations that effectively communicate categorical comparisons. This is designed as an in-class 1-hour lab.

Learning Objectives

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

• Order categorical data to maximize clarity
• Apply and customize ggplot2 themes effectively
• Add meaningful annotations to guide reader attention
• Create compound figures using patchwork

Required Packages

library(tidyverse)
library(gapminder)
library(ggrepel)
library(gghighlight)
library(patchwork)

Data

We’ll be working with the gapminder dataset, focusing on countries in the Americas.

Part 1: Ordering Categories Effectively

Task 1.1

Using the gapminder dataset filtered to year 2007 and continent “Americas”, create a dot plot showing GDP per capita by country.

Requirements:

• Filter to 2007 and Americas
• Order countries by GDP per capita (lowest to highest)
• Use geom_point() for a Cleveland dot plot
• Place country names on the y-axis
• Format GDP with scales::dollar()
• Add appropriate title and labels

Hint: Use fct_reorder() to order countries by GDP.

Task 1.2

Convert your dot plot from Task 1.1 into a lollipop chart by adding line segments from 0 to each point.

Requirements:

• Use geom_segment() with gray lines
• Keep colored points from Task 1.1
• Maintain the same ordering and labels

Question: In 1-2 sentences, explain why ordering by value is more effective than alphabetical ordering for this visualization.

Part 2: Theme Customization

Task 2.1

Creat a plot showing life expectancy over time for countries in the Americas.

Requirements:

• Filter gapminder to Americas only
• Create line plot with year on x-axis and lifeExp on y-axis
• Use geom_line() with aes(group = country)
• Color all lines the same (your choice of color)
• Apply theme_minimal()
• Add title: “Life Expectancy in the Americas”

Task 2.2

Customize the theme of your plot from Task 2.1 by modifying at least 5 theme elements.

Requirements must include:

1. Plot title (size, face, color)
2. Axis titles (size or face)
3. Panel grid (modify or remove major/minor)
4. Plot or panel background color
5. Legend position (if you add one)

Make intentional design choices that improve readability.

Task 2.3

Create a custom theme function called theme_yourname() that encapsulates your customizations from Task 2.2.

Requirements:

• Include a base_size parameter with default = 14
• Include your 5+ theme customizations from Task 2.2
• Test it by applying to your plot

Example structure:

theme_yourname <- function(base_size = 14) {
  theme_minimal(base_size = base_size) +
  theme(
    # Your customizations here
  )
}

Part 3: Annotations

Task 3.1

Using your Americas life expectancy plot from Task 2.1, add labels to identify specific countries using gghighlight.

Requirements:

• Highlight only countries with life expectancy > 75 years in 2007
• Use gghighlight() with use_direct_label = TRUE
• Keep your custom theme from Task 2.3

Task 3.2

Create an alternative version using ggforce::geom_mark_ellipse() instead of gghighlight to highlight outliers.

Requirements:

• Start with the same base Americas plot
• Identify 2-3 countries with notably low life expectancy (you can see these on the lower part of the plot)
• Use geom_mark_ellipse() to draw an ellipse around these countries
• Add a description label to the ellipse

Hint: You’ll need to filter the data to just those countries for geom_mark_ellipse()

Question: Which annotation method (Task 3.1 with gghighlight or Task 3.2 with geom_mark_ellipse) do you prefer for this data and why? Write 2-3 sentences.

Part 4: Annotations with ggrepel

Task 4.1

Create a scatter plot using gapminder data filtered to 2007 and Americas showing GDP per capita (x-axis, log scale) vs. life expectancy (y-axis).

Requirements:

• Filter to Americas and year 2007
• Use geom_point() with appropriate size
• Use scale_x_log10(labels = scales::dollar)
• Apply your custom theme

Task 4.2

Add labels to the plot from Task 4.1 using geom_text_repel().

Requirements:

• Label the 3 countries with highest GDP per capita
• Label the 3 countries with lowest life expectancy
• Use geom_text_repel() to avoid overlaps
• Differentiate these labeled countries with a different color or size

Hint: Create a variable to identify which countries should be labeled, then use that in your aesthetics and geom_text_repel().

Part 5: Compound Figures

Task 5.1

Create two plots that tell a story about the Americas:

• Plot A: Your lollipop chart from Task 1.2 (GDP per capita in 2007)
• Plot B: Your annotated scatter plot from Task 4.2 (GDP vs. life expectancy)

Combine them using patchwork side-by-side.

Requirements:

• Use the + operator to place plots side by side
• Add an overall title using plot_annotation()
• Add tags (A, B) using tag_levels
• Ensure both plots use your custom theme

Task 5.2

Experiment with the layout from Task 5.1.

Requirements:

• Try stacking the plots vertically using /
• Try making one plot larger than the other using plot_layout(widths = c(2, 1))

Question: Which layout (side-by-side, stacked, or different widths) works best for these two plots? Why?

Part 6: Saving Your Work

Task 6.1

Save your best compound figure from Part 5.

Requirements:

• Save as PNG with ggsave()
• Use 300 DPI
• Use dimensions: 10 inches wide by 6 inches tall
• Ensure white background with bg = "white"

Include the ggsave() code in your Rmd file.

Reflection Questions (No need to write- we will discuss in class next week)

Answer the following (2-3 sentences each):

1. Ordering: When visualizing categorical data, what are two situations where you would NOT want to order by value?

2. Themes: How did creating a custom theme function help with consistency? What’s one advantage of using a theme function vs. copying theme code?

3. Annotations: Compare direct labeling (like gghighlight) versus legends. What’s one situation where you’d prefer each approach?

Finishing Up

When you have finished the above, upload your rendered PDF file to Canvas.

What to Submit

• A single PDF file rendered from your R Markdown
• Include all code, plots, and written answers
• Ensure plots are clearly visible and properly sized