data-visualisation

---
name: data-visualisation
description: Choose and produce accurate charts, maps, and analytical visuals for reports or
  presentations. Use when evidence needs a defensible comparison with scales, units, uncertainty,
  sources, and accessible data.
license: MIT
metadata:
  author: org-xelius
  version: 0.2.0
  source: https://github.com/org-xelius/xelius-document-skills
---

# Data Visualisation

Choose the visual form from the analytical relationship and audience decision.

## When to use

Use for quantitative or structured qualitative evidence that benefits materially from a visual form.

## When not to use

Keep the information as prose or a table when precise lookup, not pattern recognition, is the job.

## Inputs

Clean data, data dictionary, evidence IDs, analytical question, audience, output medium, design tokens, units, periods, sample and source notes.

## Outputs

Chart specification, accessible data table, editable source, SVG, PNG, alt text, and validation notes.

## Required workflow

1. Validate the data and intended comparison.
2. Read [CHART-SELECTION.md](references/CHART-SELECTION.md).
3. State the one-message finding before selecting a chart.
4. Choose the simplest valid form; define scales, order, baseline, labels, annotations, units, and notes.
5. Apply the chart theme and output-specific density.
6. Export SVG and PNG; render and inspect both in the target report or PPTX.
7. Verify visual values against source data and provide an accessible table.

## Hard rules

- Never invent, interpolate, or silently impute missing data.
- Use no 3D or decorative forms that obscure magnitude.
- Do not truncate scales in ways that mislead; disclose non-zero baselines.
- Use direct labels where practical and consistent scales when comparison is intended.
- State units, periods, sources, samples, and relevant uncertainty.
- Give presentation charts larger labels, fewer variables, stronger focus, and simpler annotation.

## Checkpoints

Approve chart intent and form before producing a large chart set. Escalate transformations that change interpretation.

## Files to read by phase

Read chart selection before specification, then the design-system chart theme and the target medium's composition reference.

## Quality checks

Check totals, denominators, ordering, scales, ticks, colour meaning, legibility, annotation accuracy, source notes, SVG and PNG parity, and accessible-table parity.

## Failure and fallback behaviour

If data is insufficient, produce a data-gap note or an annotated table. If SVG cannot render in the target, preserve the SVG source and use a verified high-resolution PNG fallback.

## Example invocations

- "Choose the right form for changes across five regions over time."
- "Simplify this report chart for a presentation."
- "Create an accessible map and table from this dataset."

## Adjacent skills

Use table design for lookup-heavy evidence, SVG illustration for diagrams, and visual QA for rendered inspection.