AGENT SKILL

data-visualization-mastery

50+ chart types with selection criteria, dashboard composition, data tables, KPI cards, sparklines, heatmaps, and accessible data presentation, with React code (Recharts, D3, Nivo) and data-viz tokens. Use when designing charts, dashboards, or any data-dense screen.

License 许可
MIT
最近更新
2026-11-10
Stars
★ 0
Forks
0

01 前言 Front Matter

name
data-visualization-mastery
description
50+ chart types with selection criteria, dashboard composition, data tables, KPI cards, sparklines, heatmaps, and accessible data presentation, with React code (Recharts, D3, Nivo) and data-viz tokens. Use when designing charts, dashboards, or any data-dense screen.
allowed-tools
Read, Grep, Glob

02 资源文件 7

03 SKILL.md 正文

Data Visualization Mastery

Purpose

Encode data so people see patterns, outliers and trends faster than they could by reading numbers, and never mislead them. The skill applies Tufte, Few, Bertin/Cleveland-McGill and Munzner to charts, dashboards and data tables that are honest, readable and accessible.

Core rules

  • Truthful encoding. Keep the lie factor between 0.95 and 1.05. Start bar charts at 0. Avoid dual Y-axes, 3D, rainbow scales and decorative gradients.
  • Maximize data-ink (aim above 0.8). Remove chartjunk: heavy gridlines (use opacity 0.1-0.15), redundant legends (label directly), borders, excess decimals. Use small multiples instead of crowded charts.
  • Use the most accurate channel for the most important dimension. Position, then length, angle, area, saturation, hue. A bar chart beats a pie; use pies only for part-to-whole with 2-5 slices.
  • Choose the chart by purpose: compare → bar/dot; trend → line/area/sparkline; part-to-whole → 100% stacked bar, treemap, donut (2-5); distribution → histogram/box/violin; relationship → scatter/bubble; flow → Sankey/funnel; geography → choropleth; hierarchy → treemap/sunburst. Match density to the audience: executives see 5-7 data points, analysts get interaction.
  • Color. Sequential for ordered data, diverging around a meaningful midpoint, categorical capped at 8 (group the rest as "Other"). Never rely on red/green alone; add lightness, pattern or labels. Keep series colors the same across charts. Gray for context, saturated color for focus.
  • Numbers. Use tabular figures (font-variant-numeric: tabular-nums) and right-align them. Abbreviate large values ($1.2M) and show deltas with sign, color and arrow.
  • This skill owns dashboards and data tables. desktop-app-design covers desktop chrome (windows, menus, panels, shortcuts); layout-block-intelligence covers marketing page sections; ui-pattern-intelligence covers app blocks such as kanban, calendars and activity feeds.
  • Dashboards (Few). Fit the primary view on one screen, put the most important information top-left, give every number context (target, previous period, benchmark), use 5-7 charts at most, summary first and detail on demand.
  • KPI cards. One hero number, always with context, semantic delta color (an increase can be bad, e.g. churn), an optional sparkline, 3-5 per row.
  • Animate only data changes. Data stays readable within 500ms and animations respect prefers-reduced-motion.
  • Every chart has loading, empty, error and partial-data states, resizes rather than scales, has 44×44px touch targets, and has alt text stating the insight plus a data-table alternative.

Routing

Question Read
Tufte/Bertin/Few/Munzner detail; selection by purpose, data shape and audience; palettes with hex values; type scale; grid and spacing; animation and interaction; KPI anatomy; density; anti-patterns; responsive behavior; dataviz tokens and dark mode; print CSS; the full review checklist references/dataviz-foundations.md
Picking or specifying a particular chart (50+ types: when to use, data shape, max points, design specs) references/chart-type-encyclopedia.md
Dashboard layouts: KPI rows, chart grids, filters, executive/operational/analytical, responsive collapse references/dashboard-layout-patterns.md
Data tables: column types, sort, filter, pagination, selection, editing, density, accessibility, React code references/data-table-mastery.md
Dashboard IA and metric tiers, card anatomy, page shells, interaction (zoom, brush, linked views, drill), real-time refresh and streaming, loading and stale data, alerts and thresholds, personalization, export, embedding, performance; advanced tables (grouped rows, tree tables, frozen columns, virtual scrolling, inline and bulk editing, grid vs. table) references/dashboards-and-tables.md
Accessibility: alt-text patterns, colorblind-safe palettes, pattern fills, screen readers, keyboard, sonification references/dataviz-accessibility.md
Recharts code: chart wrapper with states, tooltip, legend, responsive container, 20 chart types, KPI card, print references/dataviz-code-patterns.md

Common requests

  • "I need a chart for this data": ask for the analytical purpose and the data shape (dimensions, measures, cardinality), pick the chart from the encyclopedia, then give its design specs and code from dataviz-code-patterns.md.
  • "Design a dashboard": identify the type (executive, operational, analytical), pick the layout from dashboard-layout-patterns.md, choose a chart per panel, and apply the KPI-card and responsive rules in dataviz-foundations.md.
  • "Make this chart accessible": write alt text using the chart type's pattern, add a data table alternative, check the palette with a colorblind simulation, and add keyboard navigation if the chart is interactive.
  • "Build a data table": decide column types, sorting and filtering, pick a responsive strategy, and use a native <table> with aria-sort on the sorted <th> (role="grid" only for an interactive grid with arrow-key cell navigation), using data-table-mastery.md; for grouping, trees, inline editing or spreadsheet-style entry, see dashboards-and-tables.md.

Cross-References

  • visual-design-mastery -- Color science (oklch, palette generation) applies directly to dataviz color palettes. Typography mastery (tabular figures, alignment) is critical for number-heavy UIs.
  • ui-pattern-intelligence -- Dashboard and table patterns at the component level. This skill provides the deeper dataviz-specific design intelligence.
  • accessibility-inclusive-design -- WCAG foundation. This skill extends it with dataviz-specific accessibility patterns.
  • component-patterns-code -- Component architecture. This skill provides dataviz-specific component patterns and code.
  • design-systems-architecture -- Token architecture. This skill defines dataviz-specific tokens (chart colors, gridlines, axes).
  • performance-states-patterns -- Loading, error, and empty states apply to every chart and table component.
  • cognitive-psychology-ux -- Perceptual principles (pre-attentive processing, Gestalt grouping) underpin all visual encoding decisions.

04 同仓库更多 Skills