Sunspot area by solar latitude, 1874–2013 — austere vs. beautiful
Same dataset, same analytical content. Must Tufte-fidelity look clinical?
Pre — Tufte-austere
The canonical Maunder butterfly diagram: each darker pixel marks where sunspots appeared. Monochrome on white, no chrome, full data density. Honest, dense, complete — and emotionally inert.
Post — Tufte-fidelity at full beauty
Mirrored ridgelines, one per 2° latitude band, rendered in the cool blue-cyan palette of downshifted UV observation. Same data: 1,871 Carrington rotations × 40 bands. The migration pattern, cycle structure, and relative cycle strengths — all preserved. The gradient stroke brightens at peaks (encoding magnitude redundantly with ridge height); the glow filter is tinted blue so strong cycles physically luminesce; a hint of pink-white at the very crest gives the soap-bubble iridescence of charged plasma.
Both versions are honest. Both preserve the analytical content. The austere version is what the Tufte checklist literally recommends; the beautiful version sacrifices nothing the checklist actually demands and gains an emotional register the austere version cannot reach. Tufte's principles are about honest information density, not about a specific aesthetic.
What the skill should learn
Aesthetic register is a design choice, not a violation. The Tufte checklist is necessary but not sufficient.
Color earns its ink when it carries a real channel. Here the brightness gradient redundantly encodes magnitude on the stroke (peaks brighten) — same job ridge height already does, but legible at a different glance distance.
Effects (glow, gradients, transparency) are honest when they amplify a true encoding. The glow doesn't fabricate signal — it makes a real signal more visceral.
The "Tufte aesthetic" is a safe default, not the only acceptable destination.
Avoiding chartjunk ≠ avoiding ornament. The line is between ornament that carries information and ornament that distracts from it.