Getting Started with GGenemy

Introduction

GGenemy audits your ggplot2 visualizations to ensure they are:

Accessible to colorblind users (~8% of males)
Honest and not misleading
Clear and readable

Think of it as a linter for data visualization!

Installation

# Install from GitHub
devtools::install_github("yourusername/GGenemy")

library(GGenemy)
library(ggplot2)

Quick Example

Let’s create a plot with some accessibility issues:

# A plot using red and green (problematic for colorblind users)
problematic_plot <- ggplot(mtcars, aes(wt, mpg, color = factor(cyl))) +
  geom_point(size = 2) +
  scale_color_manual(values = c("red", "green", "blue"))

problematic_plot

Now audit it:

report <- gg_audit(problematic_plot)
print(report)
#> 
#> === GGenemy Audit Report ===
#> 
#> Plot Type: GeomPoint 
#> Layers: 1 
#> 
#> CRITICAL ISSUES (2):
#>   1. [color] Red-green color combination detected (problematic for 8% of males with deuteranopia/protanopia)
#>   2. [labels] Label 'factor(cyl)' contains programming syntax
#> 
#> WARNINGS (3):
#>   1. [color] Some colors have low contrast - may be hard to distinguish
#>   2. [scales] Unusual aspect ratio (6.01) may distort perception of relationships
#>   3. [labels] Plot has no title
#> 
#> SUGGESTIONS:
#>   - Use colorblind-safe palettes: viridis::scale_color_viridis() or RColorBrewer
#>   - Consider using coord_fixed() for appropriate aspect ratio
#>   - White backgrounds work well, but ensure sufficient contrast with data elements
#>   - Add a descriptive title: labs(title = 'Your Title Here')
#>   - Clean up label: labs(colour = 'Human Readable Name')
#> 
#> Run gg_suggest_fixes() for code recommendations.

Color Accessibility

Check for color issues specifically:

gg_audit_color(problematic_plot)
#> $issues
#> [1] "Red-green color combination detected (problematic for 8% of males with deuteranopia/protanopia)"
#> 
#> $warnings
#> [1] "Some colors have low contrast - may be hard to distinguish"
#> 
#> $suggestions
#> [1] "Use colorblind-safe palettes: viridis::scale_color_viridis() or RColorBrewer"

The red-green combination is a problem! Let’s simulate how a colorblind user sees it:

# Deuteranopia (most common type of colorblindness)
gg_simulate_cvd(problematic_plot, type = "deutan")
#> Simulating deutan color vision deficiency...
#> Original colors: green, red, blue
#> Simulated colors: #EFD63A, #A39000, #003DFB

See how the red and green look very similar?

Getting Fix Suggestions

gg_suggest_fixes(problematic_plot)
#> 
#> ====================================
#>     GGenemy Fix Suggestions
#> ====================================
#> 
#> Copy and add these layers to your plot:
#> 
#> your_plot <- ggplot(...) +
#>   geom_*(...) +
#> 
#>   # COLOR FIXES
#>   # Replace with colorblind-safe palette
#>   scale_color_viridis_d(option = 'plasma') +
#>   # Use high-contrast palette
#>   scale_color_brewer(palette = 'Set1') +
#> 
#>   # SCALE FIXES
#>   # Fix aspect ratio
#>   coord_fixed(ratio = 1) +
#> 
#>   # LABEL FIXES
#>   # Clean up labels - replace with human-readable text
#>   labs(
#>     x = 'Descriptive X Label',
#>     y = 'Descriptive Y Label',
#>     color = 'Group Name'
#>   ) +
#>   # Add descriptive title
#>   labs(title = 'Your Descriptive Title Here') +
#> 
#>   theme_minimal()  # Optional: clean theme
#> 
#> ====================================
#>   Manual Fixes Required:
#> ====================================
#> 1. Replace programming syntax in labels with descriptive text
#> 2. Add a descriptive title
#> 
#> TIP: Use auto_fix = TRUE to automatically apply some fixes:
#>      fixed_plot <- gg_suggest_fixes(your_plot, auto_fix = TRUE)

Creating an Accessible Plot

Here’s an improved version:

improved_plot <- ggplot(mtcars, aes(wt, mpg, color = factor(cyl))) +
  geom_point(size = 3) +
  scale_color_viridis_d(option = "plasma") +
  labs(
    title = "Car Weight vs Fuel Efficiency",
    x = "Weight (1000 lbs)",
    y = "Miles Per Gallon",
    color = "Cylinders"
  ) +
  theme_minimal()

improved_plot

Check it:

gg_audit(improved_plot)
#> 
#> === GGenemy Audit Report ===
#> 
#> Plot Type: GeomPoint 
#> Layers: 1 
#> 
#> CRITICAL ISSUES (1):
#>   1. [color] Red-green color combination detected (problematic for 8% of males with deuteranopia/protanopia)
#> 
#> WARNINGS (3):
#>   1. [scales] Unusual aspect ratio (6.01) may distort perception of relationships
#>   2. [text] Axis text size (0.8pt) is very small and may be hard to read
#>   3. [text] Title text is smaller than recommended
#> 
#> SUGGESTIONS:
#>   - Use colorblind-safe palettes: viridis::scale_color_viridis() or RColorBrewer
#>   - Consider using coord_fixed() for appropriate aspect ratio
#>   - Increase axis text size: theme(axis.text = element_text(size = 10))
#>   - White backgrounds work well, but ensure sufficient contrast with data elements
#> 
#> Run gg_suggest_fixes() for code recommendations.

Much better! And here’s how it looks to colorblind users:

gg_simulate_cvd(improved_plot, type = "deutan")
#> Simulating deutan color vision deficiency...
#> Original colors: #CC4678FF, #0D0887FF, #F0F921FF
#> Simulated colors: #848175FF, #002085FF, #FFF13CFF

The colors are still distinguishable!

What GGenemy Checks

Color Issues

Red-green combinations
Too many colors (>7)
Low contrast colors
Rainbow palettes (perceptually non-uniform)

Scale Issues

Bar charts not starting at zero
Truncated axes
Dual y-axes
Unlabeled log scales
Inappropriate aspect ratios

Text Issues

Text too small
Overlapping labels
Missing axis labels
Long labels that need wrapping

Accessibility

Points/lines too small
Color as the only distinguishing feature
Missing titles and labels

Best Practices

Use colorblind-safe palettes: viridis, plasma, ColorBrewer
Add redundant encoding: Use both color AND shape
Label everything: Axes, legends, titles
Start bar charts at zero
Keep text readable: Size 10+ points
Test with gg_simulate_cvd()

Conclusion

GGenemy helps you create visualizations that are accessible, honest, and clear. Run gg_audit() on your plots before sharing them!

For more information, visit: https://github.com/andytai7/GGenemy