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

grouped_summary <- penguins_clean %>%
  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()

What plot type did you choose? Cleveland plot
Why is this plot appropriate for this data? It separates each species into its sex for comparison and shows the center of the data.
What patterns do you observe? There are significant differences between the bill lengths of the sex within all three species. Confidence intervals could be added to make this a forest plot which would provide even more information.

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

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

penguins_clean <- penguins_clean %>%
  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")

Which species has the highest body mass? Gentoo has the highest body mass.
Which species shows the greatest variability? It seems like both Adelie and Gentoo have pretty equal variability, with both having more variability than Chinstrap.
What does this plot show that a boxplot alone would not? This plot shows density and raw data points, which gives more information about the shape of the data than a boxplot alone.

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

# Don't forget, you will need to make summary data. 
library(ggplot2)
library(dplyr)
library(palmerpenguins)
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`.

Which group has the highest mean body mass? Gentoo penguins had the highest mean body mass.
Which group has the widest confidence interval? Why? Male chinstraps had the widest confidence interval. They likely have more variation in body mass than other species.
Do any groups appear clearly different (based on overlap?) Gentoo peguins are clearly the most different because there is not overlap in confidence intervals between male and female gentoos, but there is also no overlap with other species either.

Raincloud, Sina, Cleveland and Correlation

2026-04-16