visualizing-data

Data Visualization Component Library

Systematic guidance for selecting and implementing effective data visualizations, matching data characteristics with appropriate visualization types, ensuring clarity, accessibility, and impact.

Overview

Data visualization transforms raw data into visual representations that reveal patterns, trends, and insights. This skill provides:

1. Selection Framework: Systematic decision trees from data type + purpose → chart type
2. 24+ Visualization Methods: Organized by analytical purpose
3. Accessibility Patterns: WCAG 2.1 AA compliance, colorblind-safe palettes
4. Performance Strategies: Optimize for dataset size (<1000 to >100K points)
5. Multi-Language Support: JavaScript/TypeScript (primary), Python, Rust, Go

Quick Start Workflow

Step 1: Assess Data

What type? [categorical | continuous | temporal | spatial | hierarchical]
How many dimensions? [1D | 2D | multivariate]
How many points? [<100 | 100-1K | 1K-10K | >10K]

Step 2: Determine Purpose

What story to tell? [comparison | trend | distribution | relationship | composition | flow | hierarchy | geographic]

Step 3: Select Chart Type

Quick Selection:

• Compare 5-10 categories → Bar Chart
• Show sales over 12 months → Line Chart
• Display distribution of ages → Histogram or Violin Plot
• Explore correlation → Scatter Plot
• Show budget breakdown → Treemap or Stacked Bar

Complete decision trees: See references/selection-matrix.md

Step 4: Implement

See language sections below for recommended libraries.

Step 5: Apply Accessibility

• Add text alternative (aria-label)
• Ensure 3:1 color contrast minimum
• Use colorblind-safe palette
• Provide data table alternative

Step 6: Optimize Performance

• <1000 points: Standard SVG rendering
• >1000 points: Sampling or Canvas rendering
• Very large: Server-side aggregation

Purpose-First Selection

Match analytical purpose to chart type:

| Purpose | Chart Types | |---------|-------------| | Compare values | Bar Chart, Lollipop Chart | | Show trends | Line Chart, Area Chart | | Reveal distributions | Histogram, Violin Plot, Box Plot | | Explore relationships | Scatter Plot, Bubble Chart | | Explain composition | Treemap, Stacked Bar, Pie Chart (<6 slices) | | Visualize flow | Sankey Diagram, Chord Diagram | | Display hierarchy | Sunburst, Dendrogram, Treemap | | Show geographic | Choropleth Map, Symbol Map |

Visualization Catalog

Tier 1: Fundamental Primitives

General audiences, straightforward data stories:

• Bar Chart: Compare categories
• Line Chart: Show trends over time
• Scatter Plot: Explore relationships
• Pie Chart: Part-to-whole (max 5-6 slices)
• Area Chart: Emphasize magnitude over time

Tier 2: Purpose-Driven

Specific analytical insights:

• Comparison: Grouped Bar, Lollipop, Bullet Chart
• Trend: Stream Graph, Slope Graph, Sparklines
• Distribution: Violin Plot, Box Plot, Histogram
• Relationship: Bubble Chart, Hexbin Plot
• Composition: Treemap, Sunburst, Waterfall
• Flow: Sankey Diagram, Chord Diagram

Tier 3: Advanced

Complex data, sophisticated audiences:

• Multi-dimensional: Parallel Coordinates, Radar Chart, Small Multiples
• Temporal: Gantt Chart, Calendar Heatmap, Candlestick
• Network: Force-Directed Graph, Adjacency Matrix

Detailed descriptions: See references/chart-catalog.md

Accessibility Requirements (WCAG 2.1 AA)

Text Alternatives

<figure role="img" aria-label="Sales increased 15% from Q3 to Q4">
  <svg>...</svg>
</figure>

Color Requirements

• Non-text UI elements: 3:1 minimum contrast
• Text: 4.5:1 minimum (or 3:1 for large text ≥24px)
• Don't rely on color alone - use patterns/textures + labels

Colorblind-Safe Palettes

IBM Palette (Recommended):

#648FFF (Blue), #785EF0 (Purple), #DC267F (Magenta),
#FE6100 (Orange), #FFB000 (Yellow)

Avoid: Red/Green combinations (8% of males have red-green colorblindness)

Keyboard Navigation

• Tab through interactive elements
• Enter/Space to activate tooltips
• Arrow keys to navigate data points

Complete accessibility guide: See references/accessibility.md

Performance by Data Volume

| Rows | Strategy | Implementation | |------|----------|----------------| | <1,000 | Direct rendering | Standard libraries (SVG) | | 1K-10K | Sampling/aggregation | Downsample to ~500 points | | 10K-100K | Canvas rendering | Switch from SVG to Canvas | | >100K | Server-side aggregation | Backend processing |

JavaScript/TypeScript Implementation

Recharts (Business Dashboards)

Composable React components, declarative API, responsive by default.

npm install recharts

import { LineChart, Line, XAxis, YAxis, Tooltip, ResponsiveContainer } from 'recharts';

const data = [
  { month: 'Jan', sales: 4000 },
  { month: 'Feb', sales: 3000 },
  { month: 'Mar', sales: 5000 },
];

export function SalesChart() {
  return (
    <ResponsiveContainer width="100%" height={300}>
      <LineChart data={data}>
        <XAxis dataKey="month" />
        <YAxis />
        <Tooltip />
        <Line type="monotone" dataKey="sales" stroke="#8884d8" />
      </LineChart>
    </ResponsiveContainer>
  );
}

D3.js (Custom Visualizations)

Maximum flexibility, industry standard, unlimited chart types.

npm install d3

Plotly (Scientific/Interactive)

3D visualizations, statistical charts, interactive out-of-box.

npm install react-plotly.js plotly.js

Detailed examples: See references/javascript/

Python Implementation

Common Libraries:

• Plotly - Interactive charts (same API as JavaScript)
• Matplotlib - Publication-quality static plots
• Seaborn - Statistical visualizations
• Altair - Declarative visualization grammar

When building Python implementations:

1. Follow universal patterns above
2. Use RESEARCH_GUIDE.md to research libraries
3. Add to references/python/

Integration with Design Tokens

Reference the design-tokens skill for theming:

--chart-color-primary
--chart-color-1 through --chart-color-10
--chart-axis-color
--chart-grid-color
--chart-tooltip-bg

<Line stroke="var(--chart-color-primary)" />

Light/dark/high-contrast themes work automatically via design tokens.

Common Mistakes to Avoid

1. Chart-first thinking - Choose based on data + purpose, not aesthetics
2. Pie charts for >6 categories - Use sorted bar chart instead
3. Dual-axis charts - Usually misleading, use small multiples
4. 3D when 2D sufficient - Adds complexity, reduces clarity
5. Rainbow color scales - Not perceptually uniform, not colorblind-safe
6. Truncated y-axis - Indicate clearly or start at zero
7. Too many colors - Limit to 6-8 distinct categories
8. Missing context - Always label axes, include units

Quick Decision Tree

START: What is your data?

Categorical (categories/groups)
  ├─ Compare values → Bar Chart
  ├─ Show composition → Treemap or Pie Chart (<6 slices)
  └─ Show flow → Sankey Diagram

Continuous (numbers)
  ├─ Single variable → Histogram, Violin Plot
  └─ Two variables → Scatter Plot

Temporal (time series)
  ├─ Single metric → Line Chart
  ├─ Multiple metrics → Small Multiples
  └─ Daily patterns → Calendar Heatmap

Hierarchical (nested)
  ├─ Proportions → Treemap
  └─ Show depth → Sunburst, Dendrogram

Geographic (locations)
  ├─ Regional aggregates → Choropleth Map
  └─ Point locations → Symbol Map

References

Selection Guides:

• references/chart-catalog.md - All 24+ visualization types
• references/selection-matrix.md - Complete decision trees

Technical Guides:

• references/accessibility.md - WCAG 2.1 AA patterns
• references/color-systems.md - Colorblind-safe palettes
• references/performance.md -