ggrain: a R package for raincloud plots

The `ggrain` R package introduces the `geom_rain` function to the grammar of graphics ecosystem, enabling users to easily generate comprehensive raincloud plots that combine dot, box, and violin plots for transparent and robust data visualization.

The Gist

Imagine you are trying to describe a crowd of people to a friend.

The Old Way (The Bar Plot):
Traditionally, scientists have used a "bar plot" to show data. This is like looking at a crowd from far away and just saying, "The average height of this group is 5 feet 9 inches." You get a single number, but you have no idea if everyone is exactly that height, if there are some giants and some dwarfs, or if the group is a mix of both. It's a summary that hides the messy, interesting reality of the individual people.

The New Way (The Raincloud Plot):
The authors of this paper introduced a better way to visualize data called a Raincloud Plot. Think of this not as a single number, but as a full weather report for your data. It combines three different views into one picture:

1. The Rain (Dot Plot): Imagine tiny raindrops falling. Each drop represents one single person or data point. You can see exactly how many people there are, spot the outliers (the "hailstones" that are way bigger or smaller than the rest), and see the raw truth without any distortion.
2. The Cloud (Violin Plot): Above the rain, there is a fluffy cloud. This shows the shape of the data. Is the data bunched up in the middle like a dense cloud? Is it spread out thin like a wispy mist? This tells you how the data is distributed.
3. The Umbrella (Box Plot): Under the cloud, there is a sturdy umbrella structure. This gives you the key statistics, like the middle point (median) and the range where most of the data lives.

The Problem They Solved:
While scientists loved these "Raincloud" pictures because they were so clear and honest, they were hard to make in the popular software tool called R. It was like having a great recipe but no specific kitchen gadget to make it easily. You had to build the whole thing from scratch every time.

The Solution: ggrain
The authors created a new tool called ggrain (a package for the R programming language).

Think of ggrain as a magic kitchen appliance. Before, if you wanted to bake a "Raincloud Cake," you had to mix the batter, shape the cloud, and arrange the raindrops by hand. Now, with ggrain, you just press one button (a function called geom_rain), and it instantly builds the whole complex, beautiful plot for you.

Why is this cool?

• It's flexible: You can use it to compare different groups (like mice vs. humans, or before vs. after a treatment).
• It tells a story over time: You can connect the dots to show how the same person changed over time, like watching a seed grow into a tree, rather than just comparing two different trees.
• It's open: It's free for everyone to use, helping scientists show their data in a way that is honest, transparent, and easy for everyone to understand.

In short, this paper is about giving scientists a simple, powerful tool to stop hiding the details of their data and start showing the full, beautiful picture of what's really happening.

Technical Summary

Based on the provided preprint, here is a detailed technical summary of the paper regarding the ggrain R package.

1. Problem Statement

The paper identifies a critical gap in scientific data visualization: the over-reliance on bar plots, which obscure the underlying distributional properties of data.

• Limitations of Bar Plots: They are sensitive to distortion, fail to represent raw data accurately, and cannot display potential differences in distribution (e.g., variance or skewness). This often leads to the misinterpretation of statistical magnitudes and lacks transparency.
• The Gap in R Ecosystem: While "raincloud plots" (a hybrid visualization method) have gained widespread popularity for their ability to combine raw data, summary statistics, and distribution shapes, there was previously no R package that implemented them in alignment with the "grammar of graphics" (the foundational philosophy of the ggplot2 ecosystem). Existing solutions were either non-existent in R or required complex, non-standard coding.

2. Methodology

The authors developed ggrain, an open-source R package designed to integrate raincloud plots seamlessly into the ggplot2 framework.

• Core Function: The package centers on a single, user-friendly function: geom_rain().
• Visual Composition: The methodology combines three distinct plot types into a single, cohesive visualization:
  1. Dot Plot: Displays raw data points with minimal distortion to allow for quick assessment of sample size ($N$) and outlier identification.
  2. Box Plot: Displays key distributional summary statistics, specifically the median and interquartile range (IQR).
  3. Violin Plot: Transparently displays the kernel density estimation (underlying distribution) of the data.
• Implementation: The package leverages the modular nature of ggplot2, allowing users to layer rainclouds with standard aesthetic mappings, faceting, and statistical transformations.

3. Key Contributions

The paper highlights several technical and practical contributions of the ggrain package:

• Grammar of Graphics Alignment: It fills the void in the R ecosystem by providing a ggplot2-native implementation, ensuring consistency with existing R visualization workflows.
• Advanced Plotting Capabilities: Beyond basic plotting, ggrain supports complex experimental designs:
  • Factorial Grouping: Handling multiple categorical variables.
  • Continuous Covariates: Mapping secondary continuous variables to aesthetics.
  • Longitudinal Data: A unique feature allowing observations to be connected across multiple time waves (e.g., pre/post intervention), enabling the visualization of intra-individual change alongside group averages.
• Accessibility: The package is available on both CRAN and GitHub, with comprehensive documentation (vignettes) to guide users through advanced syntax.
• Ecosystem Integration: The authors note that the raincloud concept has been extended to other tools (Python's ptitprince and JASP), but ggrain specifically targets the R statistical community.

4. Results and Demonstrations

The paper demonstrates the utility of ggrain through two primary examples (Figure 1):

• Distributional Differences: The package effectively visualizes differences in data distribution between two groups, showing not just the mean/median but the spread and shape of the data.
• Longitudinal Analysis: In a pre/post intervention scenario, the package successfully illustrates:
  • Group-level changes: Average shifts in scores.
  • Individual-level changes: Trajectories of individual subjects connected across timepoints.
  • Variance Dynamics: The ability to detect changes in variability between groups over time (e.g., observing that variability differences between groups at Time 1 disappear after an intervention).

5. Significance

The significance of this work lies in its potential to improve the transparency and rigor of scientific communication:

• Statistical Robustness: By forcing researchers to visualize raw data and distributions simultaneously, ggrain reduces the risk of misinterpreting statistical effects that might be hidden by summary-only plots (like bar charts).
• Standardization: It provides a standardized, easy-to-use tool for the R community, lowering the barrier to entry for creating high-quality, publication-ready raincloud plots.
• Reproducibility: As an open-source package with code availability on GitHub, it promotes reproducible research practices.
• Broad Applicability: The tool is applicable across diverse fields (psychology, neuroscience, medicine, policy) for visualizing factorial designs, longitudinal studies, and complex group comparisons.

In summary, ggrain solves the problem of opaque data visualization in R by providing a robust, grammar-of-graphics-compliant tool that combines raw data, summary statistics, and distribution shapes, thereby facilitating more accurate and transparent scientific communication.