AGENT SKILL

d3-visualization

Build deterministic, verifiable data visualizations with D3.js (v6). Generate standalone HTML/SVG (and optional PNG) from local data files without external network dependencies. Use when tasks require charts, plots, axes/scales, legends, tooltips, or data-driven SVG output.

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 0
Forks
1

01 前言 Front Matter

name
d3-visualization
description
Build deterministic, verifiable data visualizations with D3.js (v6). Generate standalone HTML/SVG (and optional PNG) from local data files without external network dependencies. Use when tasks require charts, plots, axes/scales, legends, tooltips, or data-driven SVG output.

02 资源文件 4

03 SKILL.md 正文

D3.js Visualization Skill

Use this skill to turn structured data (CSV/TSV/JSON) into clean, reproducible visualizations using D3.js. The goal is to produce stable outputs that can be verified by diffing files or hashing.

When to use

Activate this skill when the user asks for any of the following:

  • “Make a chart/plot/graph/visualization”
  • bar/line/scatter/area/histogram/box/violin/heatmap
  • timelines, small multiples, faceting
  • axis ticks, scales, legends, tooltips
  • data-driven SVG output for a report or web page
  • converting data to a static SVG or HTML visualization

If the user only needs a quick table or summary, don’t use D3—use a spreadsheet or plain markdown instead.


Inputs you should expect

  • One or more local data files: *.csv, *.tsv, *.json
  • A chart intent:
    • chart type (or you infer the best type)
    • x/y fields and aggregation rules
    • sorting/filtering rules
    • dimensions (width/height) and margins
    • color rules (categorical / sequential)
    • any labeling requirements (title, axis labels, units)
  • Output constraints:
    • “static only”, “no animation”, “must be deterministic”, “offline”, etc.

If details are missing, make reasonable defaults and document them in comments near the top of the output file.


Outputs you should produce

Prefer producing all of the following when feasible:

  1. dist/chart.html — standalone HTML that renders the visualization
  2. dist/chart.svg — exported SVG (stable and diff-friendly)
  3. (Optional) dist/chart.png — if the task explicitly needs a raster image

Always keep outputs in a predictable folder (default: dist/), unless the task specifies paths.


Determinism rules (non-negotiable)

To keep results stable across runs and machines:

Data determinism

  • Sort input rows deterministically before binding to marks (e.g., by x then by category).
  • Use stable grouping order (explicit Array.from(grouped.keys()).sort()).
  • Avoid locale-dependent formatting unless fixed (use d3.format, d3.timeFormat with explicit formats).

Rendering determinism

  • No randomness: do not use Math.random() or d3-random.
  • No transitions/animations by default (transitions can introduce timing variance).
  • Fixed width, height, margin, viewBox.
  • Use explicit tick counts only when needed; otherwise rely on D3 defaults but keep domains fixed.
  • Avoid layout algorithms with non-deterministic iteration unless you control seeds/iterations (e.g., force simulation). If a force layout is required:
    • fix the tick count,
    • fix initial positions deterministically (e.g., sorted nodes placed on a grid),
    • run exactly N ticks and stop.

Offline + dependency determinism

  • Do not load D3 from a CDN.
  • Pin D3 to a specific version (default: d3@7.9.0).
  • Prefer vendoring a minified D3 bundle (e.g., vendor/d3.v7.9.0.min.js) or bundling with a lockfile.

File determinism

  • Stable SVG output:
    • Avoid auto-generated IDs that may change.
    • If you must use IDs (clipPath, gradients), derive them from stable strings (e.g., "clip-plot").
  • Use LF line endings.
  • Keep numeric precision consistent (e.g., round to 2–4 decimals if needed).

Recommended project layout

If the task doesn't specify an existing structure, use:

dist/
  chart.html        # standalone HTML with inline or linked JS/CSS
  chart.svg         # exported SVG (optional but nice)
  chart.png         # rasterized (optional)
vendor/
  d3.v7.9.0.min.js  # pinned D3 library

Interactive features (tooltips, click handlers, hover effects)

When the task requires interactivity (e.g., tooltips on hover, click to highlight):

Tooltip pattern (recommended)

  1. Create a tooltip element in HTML:
<div id="tooltip" class="tooltip"></div>
  1. Style with CSS using .visible class for show/hide:
.tooltip {
    position: absolute;
    padding: 10px;
    background: rgba(0, 0, 0, 0.8);
    color: white;
    border-radius: 4px;
    pointer-events: none;  /* Prevent mouse interference */
    opacity: 0;
    transition: opacity 0.2s;
    z-index: 1000;
}

.tooltip.visible {
    opacity: 1;  /* Show when .visible class is added */
}
  1. Add event handlers to SVG elements:
svg.selectAll('circle')
    .on('mouseover', function(event, d) {
        d3.select('#tooltip')
            .classed('visible', true)  // Add .visible class
            .html(`<strong>${d.name}</strong><br/>${d.value}`)
            .style('left', (event.pageX + 10) + 'px')
            .style('top', (event.pageY - 10) + 'px');
    })
    .on('mouseout', function() {
        d3.select('#tooltip').classed('visible', false);  // Remove .visible class
    });

Key points:

  • Use opacity: 0 by default (not display: none) for smooth transitions
  • Use .classed('visible', true/false) to toggle visibility
  • pointer-events: none prevents tooltip from blocking mouse events
  • Position tooltip relative to mouse with event.pageX/pageY

In-shape label overlay (text drawn on top of marks)

Use this pattern when a mark (circle/rect/path) must carry an in-situ textual identifier (e.g., a ticker symbol inside a bubble, a short code inside a cell). The label layer must not interfere with click/hover on the mark beneath it.

Mechanism:

  1. Append a parallel <text> selection bound to the same data array as the marks. Keep the two selections sibling layers in the SVG; draw text after marks so it renders on top.

  2. Center the label on the mark with:

    • text-anchor="middle" (horizontal centering)
    • dominant-baseline="central" (vertical centering)
  3. Keep text position synchronized with the mark. If a force simulation is used, update both selections inside the same tick handler using d.x / d.y:

const labels = svg.selectAll('text.bubble-label')
    .data(data, d => d.ticker)   // same key as the circles
    .join('text')
    .attr('class', 'bubble-label')
    .attr('text-anchor', 'middle')
    .attr('dominant-baseline', 'central')
    .attr('font-size', d => Math.max(8, d.radius * 0.5))  // scale with radius
    .text(d => d.radius >= MIN_LABEL_RADIUS ? d.ticker : ''); // skip when too small

simulation.on('tick', () => {
    bubbles.attr('cx', d => d.x).attr('cy', d => d.y);
    labels.attr('x', d => d.x).attr('y', d => d.y);
});
  1. Make the overlay pointer-transparent so clicks/hovers continue to reach the mark below. Set this either in CSS or as an attribute on the text selection:
.bubble-label {
    pointer-events: none;    /* clicks/hover pass through to the circle underneath */
    user-select: none;
}
  1. Legibility threshold: when the underlying mark is too small to host readable text, emit an empty string (as shown above) rather than clipping or overflowing. Pick a MIN_LABEL_RADIUS consistent with the chosen font-size scaling.

Scope note: this pattern is for annotation layers that identify the mark they sit on. It is not a substitute for axis tick labels, external legends, or tooltip content, and should not be applied to overlays that must themselves receive events (e.g., close buttons).

Click handlers for selection/highlighting

// Add 'selected' class on click
svg.selectAll('.bar')
    .on('click', function(event, d) {
        // Remove previous selection
        d3.selectAll('.bar').classed('selected', false);
        // Add to clicked element
        d3.select(this).classed('selected', true);
    });

CSS for highlighting:

.bar.selected {
    stroke: #000;
    stroke-width: 3px;
}

Cross-view selection (bidirectional chart ↔ table linkage)

When two or more views (e.g., a chart of marks and a tabular list of rows) represent the same entities and must stay in sync — a click in either view must highlight the matching entity in both — use a shared stable key and a pair of symmetric handlers.

Mechanism:

  1. Pick one stable key that uniquely identifies an entity across views (e.g., ticker). Both the mark selection and the row selection must be data-joined with this same key.

  2. Define two symmetric selection functions. Each one is the single source of truth for its originating view, and each updates both views by the shared key:

// Called when the user clicks a bubble in the chart.
function selectFromChart(key) {
    // Clear prior selection on BOTH views
    d3.selectAll('.bubble').classed('selected', false);
    d3.selectAll('#data-table tbody tr').classed('selected', false);
    // Set selection on BOTH views by shared key
    d3.selectAll('.bubble').filter(d => d.ticker === key).classed('selected', true);
    d3.selectAll('#data-table tbody tr').filter(d => d.ticker === key).classed('selected', true);
}

// Called when the user clicks a row in the table.
function selectFromTable(key) {
    // Same symmetric behavior — the two handlers are intentionally mirror images.
    d3.selectAll('.bubble').classed('selected', false);
    d3.selectAll('#data-table tbody tr').classed('selected', false);
    d3.selectAll('.bubble').filter(d => d.ticker === key).classed('selected', true);
    d3.selectAll('#data-table tbody tr').filter(d => d.ticker === key).classed('selected', true);
}
  1. Wire each view's click to its handler:
bubbles.on('click', (event, d) => selectFromChart(d.ticker));
rows.on('click',    (event, d) => selectFromTable(d.ticker));
  1. Style .selected on both views so the highlight is visible wherever the entity appears:
circle.bubble.selected {
    stroke: #000;
    stroke-width: 3px;
}
tr.selected {
    background: #e3f2fd;
    border-left: 3px solid #2196f3;
}
  1. If the bundled helpers are used (scripts/interactive_table_example.js exposes highlightTableRow(ticker) and the bubble-chart example exposes highlightBubble(ticker)), invoke them symmetrically from both click handlers — not just one direction — so that clicking either view updates the other.

Interaction with the in-shape label overlay: because selection relies on click events reaching the mark, any overlay drawn on top of the mark (see “In-shape label overlay”) must have pointer-events: none. Otherwise the overlay silently intercepts clicks and the bidirectional linkage will appear to work from table → chart but not from chart → table.

Scope note: this pattern applies to selection of a single shared entity. Do not reuse the same channel to synchronize unrelated state (filters, sorts, hover-only highlights, multi-select) unless those are explicitly required — those need their own, separately-named channels.

Conditional interactivity

Sometimes only certain elements should be interactive:

.on('mouseover', function(event, d) {
    // Example: Don't show tooltip for certain categories
    if (d.category === 'excluded') {
        return;  // Exit early, no tooltip
    }
    // Show tooltip for others
    showTooltip(event, d);
})

04 同仓库更多 Skills