Concepts and Plotting Using R
By the end of this document, readers will be able to:
Data visualization is the graphical representation of information and data. In epidemiology and public health, effective visualization serves multiple critical purposes: exploring data patterns, communicating findings to diverse audiences, supporting evidence-based decision making, and revealing insights that might be hidden in raw numbers.
As Edward Tufte eloquently stated in his seminal work The Visual Display of Quantitative Information (1983): “Excellence in statistical graphics consists of complex ideas communicated with clarity, precision and efficiency.”
The theoretical foundation underlying modern data visualization, particularly the ggplot2 package, is based on Leland Wilkinson’s “Grammar of Graphics” (1999). This framework treats plots as composed of distinct layers and components that can be systematically combined:
Every visualization should have a clear purpose and message. Before creating any plot, ask yourself:
Tufte’s principle emphasizes that the majority of ink should be devoted to displaying data, not decorative elements. Remove unnecessary:
Different data types require different visualization approaches:
Guide the viewer’s attention through:
Good visualizations exhibit:
Example: A well-designed epidemiological curve showing COVID-19 cases over time with clear date labels, appropriate scale, and distinct colors for different case types (confirmed, probable, deaths).
Poor visualizations often include:
Example: A pie chart with too many small slices, making it impossible to distinguish between categories, when a horizontal bar chart would be more effective.
Let’s begin by loading the necessary packages and exploring our dataset.
# Load required packages
library(tidyverse) # Includes ggplot2, dplyr, and other tidyverse packages
library(gapminder) # Contains the gapminder dataset
library(patchwork) # For combining multiple plots
library(scales) # For better axis formatting
# Set a clean theme as default
theme_set(theme_minimal())The Gapminder dataset provides longitudinal data on countries’ socioeconomic indicators. This cleaned excerpt contains observations for 142 countries from 1952 to 2007, measured every 5 years.
# Load and examine the gapminder dataset
glimpse(gapminder)Rows: 1,704
Columns: 6
$ country <fct> "Afghanistan", "Afghanistan", "Afghanistan", "Afghanistan", …
$ continent <fct> Asia, Asia, Asia, Asia, Asia, Asia, Asia, Asia, Asia, Asia, …
$ year <int> 1952, 1957, 1962, 1967, 1972, 1977, 1982, 1987, 1992, 1997, …
$ lifeExp <dbl> 28.801, 30.332, 31.997, 34.020, 36.088, 38.438, 39.854, 40.8…
$ pop <int> 8425333, 9240934, 10267083, 11537966, 13079460, 14880372, 12…
$ gdpPercap <dbl> 779.4453, 820.8530, 853.1007, 836.1971, 739.9811, 786.1134, …# Display first few rows
head(gapminder)# A tibble: 6 × 6
country continent year lifeExp pop gdpPercap
<fct> <fct> <int> <dbl> <int> <dbl>
1 Afghanistan Asia 1952 28.8 8425333 779.
2 Afghanistan Asia 1957 30.3 9240934 821.
3 Afghanistan Asia 1962 32.0 10267083 853.
4 Afghanistan Asia 1967 34.0 11537966 836.
5 Afghanistan Asia 1972 36.1 13079460 740.
6 Afghanistan Asia 1977 38.4 14880372 786.# Summary statistics
summary(gapminder) country continent year lifeExp
Afghanistan: 12 Africa :624 Min. :1952 Min. :23.60
Albania : 12 Americas:300 1st Qu.:1966 1st Qu.:48.20
Algeria : 12 Asia :396 Median :1980 Median :60.71
Angola : 12 Europe :360 Mean :1980 Mean :59.47
Argentina : 12 Oceania : 24 3rd Qu.:1993 3rd Qu.:70.85
Australia : 12 Max. :2007 Max. :82.60
(Other) :1632
pop gdpPercap
Min. :6.001e+04 Min. : 241.2
1st Qu.:2.794e+06 1st Qu.: 1202.1
Median :7.024e+06 Median : 3531.8
Mean :2.960e+07 Mean : 7215.3
3rd Qu.:1.959e+07 3rd Qu.: 9325.5
Max. :1.319e+09 Max. :113523.1
The dataset contains six variables:
Let’s start with the simplest case - plotting one variable.
# Basic histogram of life expectancy
gapminder |>
ggplot(aes(x = lifeExp)) +
geom_histogram(bins = 30, fill = "steelblue", alpha = 0.7) +
labs(
title = "Distribution of Life Expectancy",
subtitle = "Gapminder dataset (1952-2007)",
x = "Life Expectancy (years)",
y = "Frequency"
)
# Density plot for smoother distribution
gapminder |>
ggplot(aes(x = lifeExp)) +
geom_density(fill = "steelblue", alpha = 0.7) +
labs(
title = "Density Distribution of Life Expectancy",
x = "Life Expectancy (years)",
y = "Density"
)
Now let’s explore relationships between two variables.
# Basic scatter plot: GDP per capita vs Life Expectancy
gapminder |>
ggplot(aes(x = gdpPercap, y = lifeExp)) +
geom_point(alpha = 0.6) +
labs(
title = "Life Expectancy vs GDP per Capita",
x = "GDP per Capita (USD)",
y = "Life Expectancy (years)"
)
The relationship is clearer with a log transformation:
# Improved scatter plot with log scale
gapminder |>
ggplot(aes(x = gdpPercap, y = lifeExp)) +
geom_point(alpha = 0.6) +
scale_x_log10(labels = scales::label_dollar()) +
labs(
title = "Life Expectancy vs GDP per Capita (Log Scale)",
x = "GDP per Capita (USD, log scale)",
y = "Life Expectancy (years)"
)
We can encode additional variables using color, size, or shape:
# Add continent as color
gapminder |>
ggplot(aes(x = gdpPercap, y = lifeExp, color = continent)) +
geom_point(alpha = 0.7, size = 2) +
scale_x_log10(labels = scales::label_dollar()) +
labs(
title = "Life Expectancy vs GDP per Capita by Continent",
x = "GDP per Capita (USD, log scale)",
y = "Life Expectancy (years)",
color = "Continent"
)
# Add population as size
gapminder |>
filter(year == 2007) |> # Focus on most recent year
ggplot(aes(x = gdpPercap, y = lifeExp, size = pop, color = continent)) +
geom_point(alpha = 0.7) +
scale_x_log10(labels = scales::label_dollar()) +
scale_size_continuous(
name = "Population",
labels = scales::label_number(scale = 1e-6, suffix = "M"),
range = c(1, 15)
) +
labs(
title = "Life Expectancy vs GDP per Capita in 2007",
subtitle = "Point size represents population",
x = "GDP per Capita (USD, log scale)",
y = "Life Expectancy (years)",
color = "Continent"
)
# Add smoothed trend lines
gapminder |>
ggplot(aes(x = gdpPercap, y = lifeExp)) +
geom_point(alpha = 0.4) +
geom_smooth(method = "loess", se = TRUE, color = "red", linewidth = 1.2) +
scale_x_log10(labels = scales::label_dollar()) +
labs(
title = "Life Expectancy vs GDP per Capita with Trend Line",
x = "GDP per Capita (USD, log scale)",
y = "Life Expectancy (years)"
)
# Separate trend lines by continent
gapminder |>
ggplot(aes(x = gdpPercap, y = lifeExp, color = continent)) +
geom_point(alpha = 0.5) +
geom_smooth(method = "loess", se = FALSE, linewidth = 1.2) +
scale_x_log10(labels = scales::label_dollar()) +
labs(
title = "Life Expectancy vs GDP per Capita by Continent",
subtitle = "With continent-specific trend lines",
x = "GDP per Capita (USD, log scale)",
y = "Life Expectancy (years)",
color = "Continent"
)
Faceting allows us to create multiple subplots based on categorical variables.
# Facet by continent
gapminder |>
ggplot(aes(x = gdpPercap, y = lifeExp)) +
geom_point(alpha = 0.6) +
geom_smooth(method = "loess", se = FALSE, color = "red") +
scale_x_log10(labels = scales::label_dollar()) +
facet_wrap(~continent, nrow = 2) +
labs(
title = "Life Expectancy vs GDP per Capita by Continent",
x = "GDP per Capita (USD, log scale)",
y = "Life Expectancy (years)"
)
# Life expectancy trends over time by continent
gapminder |>
ggplot(aes(x = year, y = lifeExp)) +
geom_line(aes(group = country), alpha = 0.3) +
geom_smooth(method = "loess", se = TRUE, color = "red", linewidth = 1.5) +
facet_wrap(~continent, nrow = 2) +
labs(
title = "Life Expectancy Trends Over Time",
subtitle = "Individual country trajectories (gray) with continent averages (red)",
x = "Year",
y = "Life Expectancy (years)"
)
# Create categorical variables for faceting
gapminder_cat <- gapminder |>
filter(year %in% c(1952, 1977, 2002)) |>
mutate(
gdp_level = case_when(
gdpPercap < 1000 ~ "Low GDP",
gdpPercap < 10000 ~ "Medium GDP",
TRUE ~ "High GDP"
),
gdp_level = factor(gdp_level, levels = c("Low GDP", "Medium GDP", "High GDP"))
)
gapminder_cat |>
ggplot(aes(x = continent, y = lifeExp)) +
geom_boxplot(aes(fill = continent), alpha = 0.7) +
facet_grid(year ~ gdp_level) +
theme(axis.text.x = element_text(angle = 45, hjust = 1)) +
labs(
title = "Life Expectancy by Continent, GDP Level, and Year",
x = "Continent",
y = "Life Expectancy (years)",
fill = "Continent"
) +
guides(fill = "none") # Remove redundant legend
Using the patchwork package to combine different visualizations:
# Create individual plots
p1 <- gapminder |>
filter(year == 2007) |>
ggplot(aes(x = continent, y = lifeExp, fill = continent)) +
geom_boxplot(alpha = 0.7) +
labs(title = "Life Expectancy by Continent (2007)",
x = "Continent", y = "Life Expectancy (years)") +
theme(legend.position = "none")
p2 <- gapminder |>
filter(year == 2007) |>
ggplot(aes(x = continent, y = gdpPercap, fill = continent)) +
geom_boxplot(alpha = 0.7) +
scale_y_log10(labels = scales::label_dollar()) +
labs(title = "GDP per Capita by Continent (2007)",
x = "Continent", y = "GDP per Capita (USD, log scale)") +
theme(legend.position = "none")
p3 <- gapminder |>
filter(year == 2007) |>
ggplot(aes(x = gdpPercap, y = lifeExp, color = continent)) +
geom_point(size = 3, alpha = 0.8) +
scale_x_log10(labels = scales::label_dollar()) +
labs(title = "GDP vs Life Expectancy (2007)",
x = "GDP per Capita (USD, log scale)",
y = "Life Expectancy (years)",
color = "Continent")
# Combine plots using patchwork
(p1 + p2) / p3 +
plot_annotation(
title = "Global Health and Economic Indicators in 2007",
subtitle = "Data from Gapminder",
theme = theme(plot.title = element_text(size = 16, hjust = 0.5))
)
# Analyze top and bottom performers in life expectancy improvement
life_exp_change <- gapminder |>
filter(year %in% c(1952, 2007)) |>
select(country, continent, year, lifeExp) |>
pivot_wider(names_from = year, values_from = lifeExp) |>
mutate(
life_exp_change = `2007` - `1952`,
life_exp_1952 = `1952`,
life_exp_2007 = `2007`
) |>
arrange(desc(life_exp_change))
# Top 10 improvers
top_improvers <- life_exp_change |>
slice_head(n = 10)
# Bottom 10 (least improvement or decline)
bottom_improvers <- life_exp_change |>
slice_tail(n = 10)
# Visualize the changes
bind_rows(
top_improvers |> mutate(group = "Top 10 Improvers"),
bottom_improvers |> mutate(group = "Bottom 10")
) |>
mutate(country = fct_reorder(country, life_exp_change)) |>
ggplot(aes(x = country, y = life_exp_change, fill = continent)) +
geom_col(alpha = 0.8) +
facet_wrap(~group, scales = "free_x") +
coord_flip() +
labs(
title = "Life Expectancy Change from 1952 to 2007",
subtitle = "Countries with largest improvements and smallest changes",
x = "Country",
y = "Change in Life Expectancy (years)",
fill = "Continent"
)
# Calculate regional statistics by decade
regional_trends <- gapminder |>
mutate(decade = (year %/% 10) * 10) |>
group_by(continent, decade) |>
summarise(
avg_life_exp = mean(lifeExp),
avg_gdp = mean(gdpPercap),
total_pop = sum(pop),
n_countries = n_distinct(country),
.groups = "drop"
)
# Visualize regional trends
regional_trends |>
ggplot(aes(x = decade, y = avg_life_exp, color = continent)) +
geom_line(linewidth = 1.2) +
geom_point(size = 3) +
scale_x_continuous(breaks = seq(1950, 2000, 10)) +
labs(
title = "Average Life Expectancy Trends by Continent",
subtitle = "Decade-by-decade analysis from 1950s to 2000s",
x = "Decade",
y = "Average Life Expectancy (years)",
color = "Continent"
)
# Identify countries that experienced GDP decline while life expectancy improved
paradox_countries <- gapminder |>
filter(year %in% c(1952, 2007)) |>
select(country, continent, year, lifeExp, gdpPercap) |>
pivot_wider(
names_from = year,
values_from = c(lifeExp, gdpPercap),
names_sep = "_"
) |>
mutate(
life_exp_change = lifeExp_2007 - lifeExp_1952,
gdp_change = gdpPercap_2007 - gdpPercap_1952,
gdp_change_pct = (gdpPercap_2007 - gdpPercap_1952) / gdpPercap_1952 * 100
) |>
filter(life_exp_change > 0 & gdp_change < 0) |>
arrange(desc(life_exp_change))
# Visualize these paradoxical cases
if(nrow(paradox_countries) > 0) {
paradox_countries |>
ggplot(aes(x = gdp_change_pct, y = life_exp_change, color = continent)) +
geom_point(size = 4, alpha = 0.8) +
geom_text(aes(label = country), vjust = -0.5, size = 3) +
labs(
title = "Countries with Declining GDP but Improving Life Expectancy",
subtitle = "1952-2007 comparison",
x = "GDP per Capita Change (%)",
y = "Life Expectancy Improvement (years)",
color = "Continent"
)
} else {
print("No countries showed declining GDP with improving life expectancy")
}
# Create a comprehensive, publication-ready visualization
final_plot <- gapminder |>
filter(year %in% c(1952, 1977, 2007)) |>
ggplot(aes(x = gdpPercap, y = lifeExp)) +
geom_point(aes(size = pop, color = continent), alpha = 0.7) +
geom_smooth(method = "lm", se = TRUE, color = "black", linetype = "dashed") +
scale_x_log10(
labels = scales::label_dollar(),
breaks = c(500, 1000, 5000, 10000, 50000)
) +
scale_size_continuous(
name = "Population\n(millions)",
labels = scales::label_number(scale = 1e-6, accuracy = 1),
range = c(1, 12),
guide = guide_legend(override.aes = list(alpha = 1))
) +
scale_color_viridis_d(name = "Continent") +
facet_wrap(~year, labeller = label_both) +
labs(
title = "The Relationship Between Wealth and Health Across Time",
subtitle = "GDP per capita vs life expectancy, 1952-2007",
caption = "Data source: Gapminder Foundation | Point size represents population",
x = "GDP per Capita (2005 USD, log scale)",
y = "Life Expectancy at Birth (years)"
) +
theme_minimal(base_size = 12) +
theme(
plot.title = element_text(size = 16, face = "bold"),
plot.subtitle = element_text(size = 12, color = "gray40"),
plot.caption = element_text(size = 10, color = "gray50"),
legend.position = "bottom",
strip.text = element_text(face = "bold"),
panel.grid.minor = element_blank()
)
print(final_plot)
This document has covered both the theoretical foundations and practical implementation of data visualization for epidemiological and public health research. Key takeaways include:
The combination of solid theoretical understanding and practical R skills will enable you to create compelling, accurate, and actionable visualizations that support evidence-based decision-making in public health practice and research.
For additional practice and examples, explore the extensive ggplot2 documentation at https://ggplot2.tidyverse.org/ and the R Graphics Cookbook at https://r-graphics.org/.