Visualising data with ggplot2

Introduction to Global Health Data Science

Author
Affiliation

Amy Herring

Duke University
STA/GLHLTH 198 Fall 2026

Published

August 24, 2026

ggplot2 \(\in\) tidyverse

  • ggplot2 is the tidyverse’s data visualization package.
  • The structure of ggplot code can be summarized as:
ggplot(
  data = my_data,
  mapping = aes(
    x = x_var,
    y = y_var
  )
) +
  geom_xxx() +
  other_options

Data: Life expectancy from infancy

IHME data on location (mostly countries), World Bank region, binary sex, and estimated life expectancy and population in 2023 will be used to explore the men’s health gap (later we will add more years).

Dr. Abernathy of The Simpsons (aged 40)

glimpse(lifeexpwide2023)
Rows: 204
Columns: 6
$ location        <chr> "Afghanistan", "Albania", "Algeria", "A…
$ year            <dbl> 2023, 2023, 2023, 2023, 2023, 2023, 202…
$ worldbankregion <chr> "South Asia", "Europe and Central Asia"…
$ pop             <dbl> 38041757, 2854191, 43053054, 55312, 771…
$ Female          <dbl> 68.98985, 82.05155, 77.60938, 74.38214,…
$ Male            <dbl> 68.13812, 77.27016, 74.54654, 67.96911,…

Grampa Simpson (Matt Groening)

Life expectancy

Discuss

  • What does the color represent?
  • What overall pattern do you notice?
  • What additional information would help explain the scatter?

Building the plot

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point() +
  labs(
    title = "Life expectancy",
    subtitle = "2023",
    x = "Female life expectancy",
    y = "Male life expectancy",
    color = "World Bank Region"
  )

Coding out loud

Start with the lifeexpwide2023 data

ggplot(
  data = lifeexpwide2023
)

Coding out loud

Start with the lifeexpwide2023 data,
and map female life expectancy to the x-axis

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female
  )
)

Coding out loud

Start with the lifeexpwide2023 data,
map female life expectancy to the x-axis,
and map male life expectancy to the y-axis

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male
  )
)

Coding out loud

Start with the lifeexpwide2023 data,
map female life expectancy to the x-axis,
map male life expectancy to the y-axis,
and represent each observation with a point

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male
  )
) +
  geom_point()

Coding out loud

Start with the lifeexpwide2023 data,
map female life expectancy to the x-axis,
map male life expectancy to the y-axis,
represent each observation with a point,
and map region to the color of each point

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point()

Coding out loud

Start with the lifeexpwide2023 data,
map female life expectancy to the x-axis,
map male life expectancy to the y-axis,
represent each observation with a point,
map region to the color of each point,
and title the plot “Life expectancy.”

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point() +
  labs(
    title = "Life expectancy"
  )

Coding out loud

Start with the lifeexpwide2023 data,
map female life expectancy to the x-axis,
map male life expectancy to the y-axis,
represent each observation with a point,
map region to the color of each point,
title the plot “Life expectancy,” and add the subtitle “2023.”

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point() +
  labs(
    title = "Life expectancy", 
    subtitle = "2023"
  )

Coding out loud

Start with the lifeexpwide2023 data, map female life expectancy to the x-axis, map male life expectancy to the y-axis, represent each observation with a point, map region to the color of each point, title the plot “Life expectancy,” add the subtitle “2023,” and refine x and y axis labels.

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point() +
  labs(
    title = "Life expectancy", 
    subtitle = "2023",
     x = "Female life expectancy", 
     y = "Male life expectancy"
  )

Coding out loud

Start with the lifeexpwide2023 data, map female life expectancy to the x-axis, map male life expectancy to the y-axis, represent each observation with a point, map region to the color of each point, title the plot “Life expectancy,” add the subtitle “2023,” refine the axis labels, and fix the legend label.

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point() +
  labs(
    title = "Life expectancy", 
    subtitle = "2023",
     x = "Female life expectancy", 
     y = "Male life expectancy",
     color="World Bank Region"
  )

Coding out loud

Start with the lifeexpwide2023 data, map female life expectancy to the x-axis, map male life expectancy to the y-axis, represent each observation with a point, map region to the color of each point, title the plot “Life expectancy,” add the subtitle “2023,” refine the axis labels, fix the legend label, and add a caption for data source.

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point() +
  labs(
    title = "Life expectancy", 
    subtitle = "2023",
     x = "Female life expectancy", 
     y = "Male life expectancy",
     color="World Bank Region",
     caption = "Source: IHME"
  )

Argument names

Tip

You can omit the names of the first two arguments when building plots with ggplot().

ggplot(
  data = lifeexpwide2023,
  mapping = aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point()
ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point()

Where is the gap most pronounced?

lifeexpwide2023 %>%
  mutate(gap = Female - Male) %>%
  select(location, worldbankregion, Female, Male, gap) %>%
  arrange(desc(gap)) %>%
  print(n = 10)
# A tibble: 204 × 5
   location                    worldbankregion Female  Male   gap
   <chr>                       <chr>            <dbl> <dbl> <dbl>
 1 Ukraine                     Europe and Cen…   78.5  66.6 11.9 
 2 Palestine                   Middle East an…   72.0  60.7 11.3 
 3 Russian Federation          Europe and Cen…   78.8  67.6 11.1 
 4 Mongolia                    East Asia and …   78.2  68.0 10.2 
 5 United States Virgin Islan… Latin America …   81.6  71.7  9.94
 6 Latvia                      Europe and Cen…   80.6  70.7  9.91
 7 Belarus                     Europe and Cen…   79.2  69.5  9.67
 8 Lithuania                   Europe and Cen…   81.2  71.9  9.34
 9 Georgia                     Europe and Cen…   79.4  70.1  9.32
10 Estonia                     Europe and Cen…   82.9  74.2  8.68
# ℹ 194 more rows

Is women’s life expectancy always greater?

lifeexpwide2023 %>%
  filter(Male > Female) %>%
  mutate(gap = Male - Female) %>%
  select(location, worldbankregion, Female, Male, gap) %>%
  arrange(desc(gap))
# A tibble: 4 × 5
  location             worldbankregion         Female  Male   gap
  <chr>                <chr>                    <dbl> <dbl> <dbl>
1 United Arab Emirates Middle East and North …   79.6  80.2 0.594
2 Morocco              Middle East and North …   72.8  73.3 0.524
3 Bhutan               South Asia                72.4  72.9 0.462
4 Libya                Middle East and North …   70.0  70.3 0.315

Aesthetics

Aesthetics options

Commonly used characteristics of plotting characters that can be mapped to a specific variable in the data are:

  • color
  • shape
  • size
  • alpha (transparency)

Color

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    color = worldbankregion
  )
) +
  geom_point()

Shape

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    shape = worldbankregion
  )
) +
  geom_point()

Oops! Only 6 shapes are available, so this isn’t a good option for our more than 6 regions! Let’s fix that.

Shape, take 2

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    shape = worldbankregion
  )
) +
  geom_point() +  
  scale_shape_manual(values = c(16,17,15,18,3,4,8,0,1,2))

Oops! Only 6 shapes are available, so this isn’t a good option for our more than 6 regions! Let’s fix that.

Shape

Can also map to the same variable as color (helps for color vision deficiency)

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    shape = worldbankregion,
    color=worldbankregion
  )
) + geom_point() 
+  scale_shape_manual(values = c(16,17,15,18,3,4,8,0,1,2))

Size

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    color = worldbankregion,
    size = pop
  )
) +
  geom_point()

Alpha (Transparency)

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    color = worldbankregion,
    alpha = pop
  )
) +
  geom_point()

Mapping vs. setting aesthetics

Mapping

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male,
    size = pop
  )
) +
  geom_point()

Setting

ggplot(
  lifeexpwide2023,
  aes(
    x = Female,
    y = Male
  )
) +
  geom_point(size = 2)

Mapping vs. setting

  • Mapping: Determine the size, alpha, etc. of points based on the values of a variable in the data.
    • Goes into aes().
  • Setting: Determine the size, alpha, etc. of points not based on the values of a variable in the data.
    • Goes into geom_*() (this was geom_point() in the previous example, but we’ll learn about other geoms soon!).

Faceting

Faceting

  • Smaller plots that display different subsets of the data
  • Useful for exploring conditional relationships and large data

Faceting

lifeexpwide %>%
  filter(
    year %in% c(1999, 2009, 2023),
    worldbankregion %in% c(
      "North America",
      "South Asia",
      "Sub-Saharan Africa"
    )
  ) %>%
  ggplot(
    aes(
      x = Female,
      y = Male
    )
  ) +
  geom_point() +
  facet_grid(year ~ worldbankregion)

Various ways to facet

lifeexpwide %>%
  filter(
    year %in% c(1999, 2010, 2023),
    worldbankregion %in% c("North America","South Asia","Sub-Saharan Africa"
    )
  ) %>%
  ggplot(
    aes(x = Female, y = Male)
  ) +
  geom_point() +
  facet_wrap(~worldbankregion)

lifeexpwide %>%
  filter(year %in% c(1999, 2010, 2023),worldbankregion %in% c("North America","South Asia","Sub-Saharan Africa")) %>%
    ggplot(aes(x = Female,
               y = Male)) + 
  geom_point() +
  facet_grid(.~worldbankregion)

Faceting summary

  • facet_grid()
    • 2D grid
    • rows ~ cols
    • Use . for no split.
  • facet_wrap()
    • 1D ribbon wrapped according to the number of rows and columns specified or the available plotting area.

Facet and color

lifeexpwide %>%
  filter(
    year %in% c(1999, 2010, 2023),
    worldbankregion %in% c(
      "North America",
      "South Asia",
      "Sub-Saharan Africa"
    )
  ) %>%
  ggplot(
    aes(
      x = Female,
      y = Male,
      color = year
    )
  ) +
  geom_point() +
  facet_wrap(~worldbankregion)

Facet and color, no legend

lifeexpwide %>%
  filter(
    year %in% c(1999, 2010, 2023),
    worldbankregion %in% c(
      "North America",
      "South Asia",
      "Sub-Saharan Africa"
    )
  ) %>%
  ggplot(
    aes(
      x = Female,
      y = Male,
      color = year
    )
  ) +
  geom_point() +
  facet_wrap(~worldbankregion) +
  guides(color = "none")

Homework (Practice)

IMS Chapter 2

  • Problem 2
  • Problem 3
  • Problem 7
  • Problem 8
  • Problem 11
  • Problem 12