Raincloud Plots

What they are for

A raincloud plot combines:

a half violin plot or density plot
a boxplot
individual data points

This makes it useful for showing:

the distribution of values
the summary statistics
the raw observations

Use a raincloud plot when you want to compare the distribution of a numeric variable across groups.

# Install if needed:
# install.packages(c("ggplot2", "ggdist", "palmerpenguins", "dplyr"))

library(ggplot2)
library(ggdist)
library(palmerpenguins)

## 
## Attaching package: 'palmerpenguins'

## The following objects are masked from 'package:datasets':
## 
##     penguins, penguins_raw

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

penguins_clean <- penguins %>%
  select(species, body_mass_g) %>%
  na.omit()

ggplot(penguins_clean, aes(x = species, y = body_mass_g, fill = species)) +
  stat_halfeye(
    adjust = 0.5,
    width = 0.6,
    justification = -0.2,
    .width = 0,
    point_colour = NA
  ) +
  geom_boxplot(
    width = 0.12,
    outlier.shape = NA,
    alpha = 0.5
  ) +
  geom_jitter(
    width = 0.08,
    alpha = 0.5,
    size = 1.5
  ) +
  labs(
    title = "Raincloud Plot of Penguin Body Mass",
    x = "Species",
    y = "Body Mass (g)"
  ) +
  theme_minimal() +
  theme(legend.position = "none")

If you have a second category you want to show….

library(ggplot2)
library(ggdist)
penguins_clean <- penguins %>%
  select(species, sex, body_mass_g) %>%
  na.omit()

ggplot(penguins_clean, aes(x = species, y = body_mass_g, fill = sex)) +
  
  # half-eye distribution
  stat_halfeye(
    position = position_dodge(width = 0.75),
    adjust = 0.6,
    width = 0.55,
    .width = 0,
    justification = -0.2,
    point_colour = NA,
    alpha = 0.5
  ) +
  
  # boxplot summary
  geom_boxplot(
    aes(color = sex),
    width = 0.12,
    position = position_dodge(width = 0.75),
    outlier.shape = NA,
    alpha = 0.65,
    linewidth = 0.5
  ) +
  
  # raw data points
  geom_jitter(
    aes(color = sex),
    position = position_jitterdodge(
      jitter.width = 0.08,
      dodge.width = 0.75
    ),
    size = 1.8,
    alpha = 0.25
  ) +
  
  labs(
    title = "Penguin Body Mass by Species and Sex",
    subtitle = "Raincloud plot showing distribution, summary statistics, and individual observations",
    x = "Species",
    y = "Body Mass (g)",
    fill = "Sex",
    color = "Sex"
  ) +
  
  scale_fill_manual(values = c("female" = "mistyrose3", "male" = "darkseagreen3")) +
  scale_color_manual(values = c("female" = "indianred3", "male" = "seagreen4")) +
  
  theme_classic(base_size = 13) +
  theme(
    plot.title = element_text(face = "bold", hjust = 0.5),
    plot.subtitle = element_text(hjust = 0.5),
    axis.title = element_text(face = "bold"),
    legend.position = "right"
  )

library(ggplot2)
library(ggdist)
library(palmerpenguins)
library(dplyr)

penguins_clean <- penguins %>%
  filter(!is.na(species), !is.na(body_mass_g)) %>%
  mutate(species = factor(species, levels = c("Adelie", "Chinstrap", "Gentoo")))

ggplot(penguins_clean, aes(x = species, y = body_mass_g)) +
  
  # half violin (raincloud shape)
  stat_halfeye(
    adjust = 0.6,
    width = 0.6,
    .width = 0,
    justification = -0.3,
    point_colour = NA,
    fill = "#74a9cf",
    alpha = 0.7
  ) +
  
  # boxplot
  geom_boxplot(
    width = 0.12,
    outlier.shape = NA,
    fill = "white",
    color = "black",
    linewidth = 0.8
  ) +
  
  # points (aligned dots instead of jitter chaos)
  geom_dotplot(
    binaxis = "y",
    stackdir = "down",
    dotsize = 0.6,
    fill = "gray40",
    alpha = 0.7
  ) +
  
  labs(
    title = "Penguin Body Mass by Species",
    x = "Species",
    y = "Body Mass (g)"
  ) +
  
  theme_classic(base_size = 16) +
  theme(
    plot.title = element_text(face = "bold", hjust = 0.5),
    axis.title = element_text(face = "bold")
  )

## Bin width defaults to 1/30 of the range of the data. Pick better value with
## `binwidth`.

Sina Plots

A sina plot is similar to a jitter plot, but the points are spread based on the density of the data. That means:

where there are more points, the plot becomes wider
where there are fewer points, it stays narrow

This makes it a nice alternative to:

strip plots
jitter plots
violin plots

It shows both individual observations and distribution shape.

Use a sina plot when you want to show raw data points without them overlapping too much, while also giving a sense of density.

Points are not randomly scattered Wider sections indicate a greater concentration of values

# Install if needed:
# install.packages(c("ggplot2", "ggforce", "palmerpenguins", "dplyr"))

library(ggplot2)
library(ggforce)
library(palmerpenguins)
library(dplyr)

penguins_clean <- penguins %>%
  select(species, flipper_length_mm) %>%
  na.omit()

ggplot(penguins_clean, aes(x = species, y = flipper_length_mm, color = species)) +
  geom_sina(alpha = 0.7, size = 2) +
  labs(
    title = "Sina Plot of Penguin Flipper Length",
    x = "Species",
    y = "Flipper Length (mm)"
  ) +
  theme_minimal() +
  theme(legend.position = "none")

Cleveland Plots

A Cleveland dot plot is used to compare values across categories using dots instead of bars.

It is helpful because:

it avoids the visual bulk of barplots
it makes it easier to compare positions along a shared axis
it works especially well when there are many groups

Use a Cleveland dot plot when comparing one summary value across categories.

Examples:

average expression by gene
average body mass by species
percentage of students by major

This example compares the average body mass of penguin species.

# Install if needed:
# install.packages(c("ggplot2", "palmerpenguins", "dplyr"))

library(ggplot2)
library(palmerpenguins)
library(dplyr)

species_summary <- penguins %>%
  group_by(species) %>%
  summarise(mean_body_mass = mean(body_mass_g, na.rm = TRUE)) %>%
  arrange(mean_body_mass)

ggplot(species_summary, aes(x = mean_body_mass, y = reorder(species, mean_body_mass))) +
  geom_point(size = 4) +
  labs(
    title = "Cleveland Dot Plot of Mean Penguin Body Mass",
    x = "Mean Body Mass (g)",
    y = "Species"
  ) +
  theme_minimal()

If you have groups…

grouped_summary <- penguins %>%
  group_by(species, sex) %>%
  summarise(mean_bill_length = mean(bill_length_mm, na.rm = TRUE), .groups = "drop")

ggplot(grouped_summary, aes(x = mean_bill_length, y = species, color = sex)) +
  geom_point(size = 3, position = position_dodge(width = 0.4)) +
  labs(
    title = "Grouped Cleveland Dot Plot of Mean Bill Length",
    x = "Mean Bill Length (mm)",
    y = "Species",
    color = "Sex"
  ) +
  theme_minimal()

Forest Plots

A forest plot shows:

a point estimate (like a mean or effect size)
a confidence interval (uncertainty around that estimate)

Use a forest plot when:

you are comparing estimates across groups
you want to include uncertainty

library(ggplot2)
library(dplyr)
library(palmerpenguins)

summary_data <- penguins %>%
  group_by(species) %>%
  summarise(
    mean_mass = mean(body_mass_g, na.rm = TRUE),
    sd = sd(body_mass_g, na.rm = TRUE),
    n = n(),
    se = sd / sqrt(n),
    lower = mean_mass - 1.96 * se,
    upper = mean_mass + 1.96 * se
  )


ggplot(summary_data, aes(x = mean_mass, y = reorder(species, mean_mass))) +
  geom_point(size = 4) +
  geom_errorbarh(aes(xmin = lower, xmax = upper), height = 0.2) +
  labs(
    title = "Forest Plot of Mean Penguin Body Mass",
    x = "Mean Body Mass (g) with 95% CI",
    y = "Species"
  ) +
  theme_minimal()

## Warning: `geom_errorbarh()` was deprecated in ggplot2 4.0.0.
## ℹ Please use the `orientation` argument of `geom_errorbar()` instead.
## This warning is displayed once per session.
## Call `lifecycle::last_lifecycle_warnings()` to see where this warning was
## generated.

## `height` was translated to `width`.

If you have grouped data…

summary_grouped <- penguins %>%
  group_by(species, sex) %>%
  summarise(
    mean_mass = mean(body_mass_g, na.rm = TRUE),
    sd = sd(body_mass_g, na.rm = TRUE),
    n = n(),
    se = sd / sqrt(n),
    lower = mean_mass - 1.96 * se,
    upper = mean_mass + 1.96 * se,
    .groups = "drop"
  )

ggplot(summary_grouped, aes(x = mean_mass, y = species, color = sex)) +
  geom_point(position = position_dodge(width = 0.5), size = 3) +
  geom_errorbarh(
    aes(xmin = lower, xmax = upper),
    position = position_dodge(width = 0.5),
    height = 0.2
  ) +
  labs(
    title = "Forest Plot of Body Mass by Species and Sex",
    x = "Mean Body Mass (g) with 95% CI",
    y = "Species",
    color = "Sex"
  ) +
  theme_minimal()

## `height` was translated to `width`.

Homework

You will be using this dataset for the homework

library(ggplot2)
library(dplyr)
library(palmerpenguins)

penguins_clean <- penguins %>%
  filter(!is.na(species), !is.na(sex), !is.na(body_mass_g))

Part 1. Create one plot of your choice to visualize the relationship between:

Species
Mass
Sex

ggplot(penguins_clean, aes(x = sex, y = body_mass_g, fill = species)) +
  stat_halfeye(
    position = position_dodge(width = 0.75),
    adjust = 0.6,
    width = 0.55,
    .width = 0,
    justification = -0.2,
    point_colour = NA,
    alpha = 0.5
  ) +
  geom_boxplot(
    aes(color = species),
    width = 0.12,
    position = position_dodge(width = 0.75),
    outlier.shape = NA,
    alpha = 0.65,
    linewidth = 0.5
  ) +
  geom_jitter(
    aes(color = species),
    position = position_jitterdodge(
      jitter.width = 0.08,
      dodge.width = 0.75
    ),
    size = 1.8,
    alpha = 0.25
  ) +
  labs(
    title = "Penguin Body Mass by Sex of Species",
    subtitle = "Raincloud plot showing distribution, summary statistics, and individual observations",
    x = "Sex",
    y = "Body Mass (g)",
    fill = "Species",
    color = "Species"
  ) +
  theme_bw()

What plot type did you choose?

Raincloud plot

Why is this plot appropriate for this data?

It shows the distribution of the data, the summary statistics, and the counts so you can have a more comprehensive view of the data. This is helpful when you are looking at multiple variables

What patterns do you observe?

Male penguins are generally larger than female penguins, and the gentoo penguins are larger overall

Part 2. Create a raincloud plot showing body mass across species.

ggplot(penguins_clean, aes(x = species, y = body_mass_g)) +
  stat_halfeye(
    position = position_dodge(width = 0.75),
    adjust = 0.6,
    width = 0.55,
    .width = 0,
    justification = -0.2,
    point_colour = NA,
    alpha = 0.5
  ) +
  geom_boxplot(
    aes(color = species),
    width = 0.12,
    position = position_dodge(width = 0.75),
    outlier.shape = NA,
    alpha = 0.65,
    linewidth = 0.5
  ) +
  geom_jitter(
    aes(color = species),
    position = position_jitterdodge(
      jitter.width = 0.08,
      dodge.width = 0.75
    ),
    size = 1.8,
    alpha = 0.25
  ) +
  labs(
    title = "Penguin Body Mass by Species",
    subtitle = "Raincloud plot showing distribution, summary statistics, and individual observations",
    x = "Species",
    y = "Body Mass (g)"
  ) +
  theme_bw()+
  theme(legend.position = "NONE")

Which species has the highest body mass?

Gentoo

Which species shows the greatest variability?

Chinstrap

What does this plot show that a boxplot alone would not?

Where there are clusters of data and the overall distribution

Part 3. Create a forest plot. Now summarize the data and visualize uncertainty.

# Don't forget, you will need to make summary data.
data <- penguins %>%
  group_by(species) %>%
  summarise(
    mean_mass = mean(body_mass_g, na.rm = TRUE),
    sd = sd(body_mass_g, na.rm = TRUE),
    n = n(),
    se = sd / sqrt(n),
    lower = mean_mass - 1.96 * se,
    upper = mean_mass + 1.96 * se
  )


ggplot(data, aes(x = mean_mass, y = reorder(species, mean_mass))) +
  geom_point(size = 4) +
  geom_errorbarh(aes(xmin = lower, xmax = upper), height = 0.2) +
  labs(
    title = "Forest Plot of Mean Penguin Body Mass",
    x = "Mean Body Mass (g) with 95% CI",
    y = "Species"
  ) +
  theme_bw()

## `height` was translated to `width`.

Which group has the highest mean body mass?