reimagine-it-infographic

Content-derived infographic poster — reads the source file and builds a statistical poster from the facts already in it. Not a dashboard, not a CV, not an AntV template. Uses common-scale encodings (C

SKILL.md

GitHub

/reimagine-it infographic

Parent: ../SKILL.md — the full Content-Derived Design engine. Research pack: ../references/research/infographic-craft.md — 15-source infographic method (Cleveland–McGill, Tufte, Bertin, ISOTYPE, Minard/Snow, Cairo, Lupi, FT Visual Vocabulary, USWDS, WCAG charts, InfoAlign layouts). Domain pack: ../references/domains/infographic.md — the infographic domain non-negotiables.

Builds a statistical poster of facts already in the source — not an ops dashboard, not a CV, not an AntV template clone.

Hard non-negotiables

  1. Cleveland–McGill perceptual ranking. Position along a common scale > length > angle > area > volume > color saturation. Encode the most important comparisons as positions, not pies.
  2. ISOTYPE unit counts when the data is discrete countable things (Neurath: one pictogram = N units, repeated, not scaled).
  3. Custom glyphs from the source. Lupi/Fragapane pattern — draw a visual alphabet from the data itself (leaves, braids, snakes, whatever the content demands), not generic bar/pie/donut.
  4. Lossless data table. Every chart has a companion table; the infographic is a supplement to the data, not a replacement.
  5. No pies, donuts, gauges, or 3D charts. Pie charts rank last in Cleveland–McGill; donuts are pies with the center punched out (worse); 3D distorts perception.
  6. No fabricated KPIs. Every number on the poster comes from the source, not extrapolated or invented.

Structure router

Before drawing, sniff the source and lock a structure:

Structure When InfoAlign scenery
list Unordered peers grid, star
sequence Steps / dated beats portrait, landscape
hierarchy Tree / nesting portrait
compare Two+ groups portrait-grid
values Magnitudes portrait-grid
relation Nodes + edges the source names spiral

Do not default to gold Texas (sequence+compare on portrait-grid). Derive from THIS source.

Procedure

  1. Load ../references/research/infographic-craft.md and ../references/domains/infographic.md.
  2. Extract anchors from the source (3–5 concrete nouns, dates, numbers).
  3. Route the structure (list/sequence/hierarchy/compare/values/relation).
  4. Pick InfoAlign scenery (grid/star/portrait/landscape/portrait-grid/spiral).
  5. Derive palette from source content (≤ 5 colors).
  6. Build a custom visual alphabet from one datapoint.
  7. Encode comparisons using common-scale positions (not pies).
  8. Include a lossless data table below the poster.
  9. Verify: every number traceable to source, no fabricated KPIs, no AntV template chrome.

Must not

  • Clone gold Texas infographic's scenery or poster chrome onto a different source
  • Use a pie chart, donut, gauge, or 3D chart
  • Invent statistics the source does not contain
  • Use AntV/Looker/Datawrapper templates
  • Skip the structure router and default to portrait-grid
  • Ship without a lossless data table