Grouped Stacked Bar Charts

Complete reference for grouped stacked bar chart visualizations in RING-5.

Overview

Grouped stacked bar charts combine grouping and stacking for complex multi-dimensional comparisons. Most advanced bar chart type.

Configuration

Required:

• X-axis: Primary category (benchmark)
• Y-axis: Numeric value (cycles, bytes)
• Group by: Secondary category (config)
• Stack by: Components (operation_type)

Use Cases

Complex Performance Analysis

Compare configurations with operation breakdown:

X: benchmark
Y: cycles
Group by: config
Stack by: operation
# Groups configs, stacks operations per config

Resource Allocation

Show resource distribution across scenarios:

X: workload
Y: memory_bytes
Group by: allocation_policy
Stack by: memory_type

When to Use

Use when:

• Need both grouping and composition
• Comparing complex multi-dimensional data
• Audience is technical

Avoid when:

• Simpler charts suffice

• 4 groups or > 4 stacks

• Audience unfamiliar with chart type

Styling

Colors

• Consistent across groups
• Distinct for stack components
• Colorblind-friendly

Layout

• Clear spacing between groups
• Stack order consistent
• Comprehensive legend

Best Practices

1. Limit to 3-4 groups
2. Limit to 3-4 stack components
3. Use descriptive legend
4. Consider alternatives first
5. Add percentage mode if needed

Alternative Approaches

Instead of grouped stacked bars, consider:

• Multiple stacked bar charts (one per group)
• Faceted visualization
• Heat map for many dimensions
• Separate plots for clarity

Next Steps

• Bar Charts: Bar-Charts.md
• Creating Plots: ../Creating-Plots.md