Ridge Plot Visualization with jggridges

library(ClinicoPath)
library(ggplot2)
library(dplyr)

Introduction

The jggridges function provides comprehensive ridge plot visualization capabilities for comparing distributions across multiple groups. Ridge plots (also known as joy plots) are particularly effective for visualizing the distribution of a continuous variable across different categories, making them ideal for clinical research applications.

Ridge plots are superior to traditional methods like box plots or violin plots when:

Multiple Groups: You need to compare distributions across many groups (>5)
Distribution Shape: The shape of the distribution is important, not just summary statistics
Space Efficiency: You need to display many distributions in a compact format
Pattern Recognition: You want to identify trends or patterns across ordered categories
Visual Appeal: You need publication-quality, aesthetically pleasing visualizations

Key Features

The jggridges function supports:

Multiple Plot Types: Basic ridgelines, density ridges, gradient density ridges, and violin ridges
Flexible Aesthetics: Customizable colors, transparency, and scaling
Statistical Overlays: Quantile lines, mean lines, and jittered data points
Advanced Options: Multiple bandwidth methods, theme styles, and faceting
Statistical Output: Summary statistics and interpretations for each group
Publication Ready: Professional themes and customizable labels

Basic Ridge Plots

Simple Distribution Comparison

Let’s start with a basic ridge plot comparing age distributions across tumor grades:

# Load histopathology data
data(histopathology, package = "ClinicoPath")

# Basic ridge plot of age by tumor grade
basic_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  plot_type = "density_ridges",
  plot_title = "Age Distribution by Tumor Grade",
  x_label = "Age (years)",
  y_label = "Tumor Grade"
)

# The plot will be displayed in jamovi interface

Understanding the Output

Ridge plots display: - X-axis: The continuous variable being analyzed - Y-axis: Categories or groups for comparison - Ridge Height: Probability density at each point - Overlap: Shows how distributions overlap between groups

Plot Types

Density Ridges

The most common type, showing smooth density estimates:

# Density ridges with custom scaling
density_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  plot_type = "density_ridges",
  scale = 1.5,  # Increase ridge height
  alpha = 0.8,  # Slight transparency
  plot_title = "Density Ridge Plot of Age by Grade"
)

Gradient Density Ridges

Add color gradients to show density variation:

# Gradient density ridges
gradient_ridge <- jggridges(
  data = histopathology,
  x_var = "MeasurementA",
  y_var = "TStage",
  plot_type = "density_ridges_gradient",
  color_palette = "viridis",
  plot_title = "Gradient Ridge Plot: Biomarker A by T-Stage"
)

Basic Ridgelines

For pre-computed densities or simple line plots:

# Basic ridgeline plot
ridgeline <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  plot_type = "ridgeline",
  plot_title = "Basic Ridgeline Plot"
)

Violin Ridges

Combination of ridge plots and violin plots:

# Density ridge plot (violin_ridges not available in ggridges)
density_ridge <- jggridges(
  data = histopathology,
  x_var = "OverallTime",
  y_var = "Group",
  plot_type = "density_ridges",
  plot_title = "Density Ridge Plot: Survival Time by Treatment Group"
)

Statistical Overlays

Quantile Lines

Show specific quantiles within each distribution:

# Ridge plot with quantile lines
quantile_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  plot_type = "density_ridges",
  quantile_lines = TRUE,
  quantiles = "0.25, 0.5, 0.75",
  plot_title = "Age Distribution with Quartiles"
)

Mean Lines

Add vertical lines showing group means:

# Ridge plot with mean lines
mean_ridge <- jggridges(
  data = histopathology,
  x_var = "MeasurementB",
  y_var = "Grade_Level",
  plot_type = "density_ridges",
  show_mean = TRUE,
  plot_title = "Biomarker B Distribution with Mean Lines"
)

Jittered Points

Show individual data points below ridges:

# Ridge plot with jittered points (raincloud style)
points_ridge <- jggridges(
  data = histopathology[sample(nrow(histopathology), 100), ],  # Subset for clarity
  x_var = "Age",
  y_var = "Sex",
  plot_type = "density_ridges",
  jittered_points = TRUE,
  point_alpha = 0.4,
  plot_title = "Age Distribution by Sex with Raw Data"
)

Aesthetic Customization

Color Variables

Use a third variable to color ridges:

# Ridge plot with color mapping
color_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "TStage",
  color_var = "LVI",
  color_palette = "Set1",
  plot_type = "density_ridges",
  plot_title = "Age by T-Stage, Colored by LVI Status"
)

Scale and Transparency

Control ridge height and opacity:

# Customized scale and transparency
custom_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  scale = 2.0,      # Double the ridge height
  alpha = 0.6,      # 60% opacity
  rel_min_height = 0.02,  # Minimum height threshold
  plot_title = "Customized Ridge Heights and Transparency"
)

Theme Styles

Apply different ggplot2 themes:

# Ridge plot with minimal theme
minimal_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  theme_style = "theme_minimal",
  plot_title = "Minimal Theme Ridge Plot"
)

# Ridge plot with classic theme
classic_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  theme_style = "theme_classic",
  plot_title = "Classic Theme Ridge Plot"
)

Advanced Features

Bandwidth Selection

Control density estimation smoothness:

# Different bandwidth methods
bandwidth_methods <- c("nrd0", "nrd", "ucv", "SJ")

for (method in bandwidth_methods) {
  ridge_plot <- jggridges(
    data = histopathology,
    x_var = "Age",
    y_var = "Grade_Level",
    plot_type = "density_ridges",
    bandwidth = method,
    plot_title = paste("Bandwidth Method:", method)
  )
}

Faceting

Create separate panels for subgroups:

# Faceted ridge plot
facet_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  facet_var = "Sex",
  plot_type = "density_ridges",
  plot_title = "Age Distribution by Grade, Faceted by Sex"
)

Reverse Order

Reverse the order of groups on Y-axis:

# Reversed order ridge plot
reverse_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  reverse_order = TRUE,
  plot_title = "Reversed Order: Low Grade at Top"
)

Clinical Research Applications

Biomarker Analysis

Compare biomarker distributions across disease stages:

# Biomarker comparison across stages
biomarker_ridge <- jggridges(
  data = histopathology,
  x_var = "MeasurementA",
  y_var = "TStage",
  plot_type = "density_ridges_gradient",
  color_palette = "plasma",
  quantile_lines = TRUE,
  quantiles = "0.5",
  plot_title = "Biomarker A Levels by Tumor Stage",
  x_label = "Biomarker A Expression",
  y_label = "T Stage",
  show_statistics = TRUE
)

Treatment Response

Visualize continuous outcomes across treatment groups:

# Treatment response visualization
treatment_ridge <- jggridges(
  data = histopathology,
  x_var = "OverallTime",
  y_var = "Group",
  color_var = "Death",
  plot_type = "density_ridges",
  show_mean = TRUE,
  plot_title = "Survival Time Distribution by Treatment Group",
  x_label = "Overall Survival Time (months)",
  y_label = "Treatment Group"
)

Multi-category Comparisons

Compare distributions across multiple categorical variables:

# Complex multi-category comparison
complex_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  color_var = "LVI",
  facet_var = "Sex",
  plot_type = "density_ridges",
  scale = 1.5,
  alpha = 0.7,
  quantile_lines = TRUE,
  quantiles = "0.25, 0.5, 0.75",
  plot_title = "Comprehensive Age Distribution Analysis",
  plot_subtitle = "By Grade, LVI Status, and Sex"
)

Color Palettes

Viridis Family

Perceptually uniform color scales:

# Viridis palette options
viridis_palettes <- c("viridis", "plasma", "inferno", "magma")

for (palette in viridis_palettes) {
  ridge_plot <- jggridges(
    data = histopathology,
    x_var = "Age",
    y_var = "Grade_Level",
    plot_type = "density_ridges_gradient",
    color_palette = palette,
    plot_title = paste("Color Palette:", palette)
  )
}

ColorBrewer Palettes

Qualitative color schemes for categorical data:

# ColorBrewer palette examples
brewer_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  color_var = "Sex",
  color_palette = "Set2",
  plot_type = "density_ridges",
  plot_title = "ColorBrewer Set2 Palette"
)

Statistical Analysis

Summary Statistics

Generate detailed statistics for each group:

# Ridge plot with statistics
stats_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  plot_type = "density_ridges",
  show_statistics = TRUE,
  show_interpretation = TRUE,
  plot_title = "Age Distribution with Statistical Summary"
)

# Statistics table includes:
# - N (sample size)
# - Mean and SD
# - Median and quartiles
# - Min and Max values

Interpretation Output

Get automated interpretation of results:

# Ridge plot with interpretation
interpret_ridge <- jggridges(
  data = histopathology,
  x_var = "MeasurementA",
  y_var = "TStage",
  show_interpretation = TRUE,
  plot_title = "Biomarker Analysis with Interpretation"
)

# Interpretation includes:
# - Dataset overview
# - Overall distribution characteristics
# - Group comparisons
# - Visual insights

Best Practices

Data Preparation

Check Data Types: Ensure X variable is numeric and Y variable is categorical
Handle Missing Data: Function removes incomplete cases automatically
Group Sizes: Ensure adequate sample size per group (n > 10 recommended)
Outliers: Consider impact on density estimation

Visual Design

Scale Selection:
- Use scale = 1.0 for minimal overlap
- Use scale > 1.5 for emphasized differences
- Use scale < 1.0 for many groups
Color Choice:
- Use gradient for continuous density representation
- Use categorical colors for group comparisons
- Consider colorblind-friendly palettes
Label Clarity:
- Use descriptive axis labels
- Add informative titles and subtitles
- Consider rotating long group labels

Statistical Considerations

Bandwidth Selection:
- “nrd0” (default): Good general purpose
- “SJ”: Better for multimodal distributions
- “ucv”: Unbiased cross-validation
Quantile Display:
- Show median (0.5) for central tendency
- Show quartiles (0.25, 0.75) for spread
- Consider deciles for detailed analysis

Troubleshooting

Common Issues

Too Many Groups

# If you have many groups, adjust scale and consider ordering
many_groups_ridge <- jggridges(
  data = data_with_many_groups,
  x_var = "value",
  y_var = "group",
  scale = 0.9,  # Reduce overlap
  rel_min_height = 0.01,  # Show smaller ridges
  expand_panel = FALSE  # Compact display
)

Sparse Data

# For sparse data, adjust bandwidth
sparse_ridge <- jggridges(
  data = sparse_data,
  x_var = "value",
  y_var = "group",
  bandwidth = "SJ",  # Adaptive bandwidth
  plot_type = "density_ridges"
)

Overlapping Labels

# Handle overlapping Y-axis labels
label_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  theme_style = "theme_minimal",
  expand_panel = TRUE
)

Complete Example

Here’s a comprehensive example using all major features:

# Comprehensive ridge plot analysis
comprehensive_ridge <- jggridges(
  data = histopathology,
  x_var = "Age",
  y_var = "Grade_Level",
  color_var = "LVI",
  plot_type = "density_ridges_gradient",
  scale = 1.8,
  alpha = 0.8,
  rel_min_height = 0.01,
  bandwidth = "nrd0",
  quantile_lines = TRUE,
  quantiles = "0.1, 0.25, 0.5, 0.75, 0.9",
  show_mean = TRUE,
  color_palette = "viridis",
  reverse_order = FALSE,
  expand_panel = TRUE,
  theme_style = "theme_ridges",
  x_label = "Patient Age (years)",
  y_label = "Tumor Grade Category",
  plot_title = "Age Distribution Analysis by Tumor Grade",
  plot_subtitle = "Colored by Lymphovascular Invasion Status",
  show_statistics = TRUE,
  show_interpretation = TRUE
)

Conclusion

The jggridges function provides a powerful and flexible tool for creating ridge plots in clinical research. Key benefits include:

Efficient Visualization: Compare many distributions in a single plot
Pattern Recognition: Easily identify trends across ordered categories
Statistical Integration: Built-in summary statistics and interpretations
Publication Quality: Professional themes and customization options
Clinical Relevance: Designed for medical research applications

Ridge plots are particularly valuable when traditional methods like box plots fail to capture the full complexity of your data distributions, making them an essential tool for modern clinical data visualization.

References

Wilke, C.O. (2017). ggridges: Ridgeline plots in ggplot2. R package version 0.5.3.
Hintze, J. L., & Nelson, R. D. (1998). Violin plots: a box plot-density trace synergism. The American Statistician, 52(2), 181-184.
Kampstra, P. (2008). Beanplot: A boxplot alternative for visual comparison of distributions. Journal of Statistical Software, 28(1), 1-9.

Creating Distribution Comparisons Across Groups

ClinicoPath

2025-07-13