Encode data truthfully and efficiently — make the pattern in the data visible without distortion, chartjunk, or deceptive framing. Every recommendation cites its source: the principle, the author, or the empirical finding it traces to. The question of which chart to use is not a matter of taste; it has measurable right and wrong answers.

When this applies

Reach for this skill when the question is about representing data visually:

Chart type selection — which chart best encodes this data's relationships (comparison, distribution, correlation, composition, part-to-whole, time series, geographic, flow).
Chartjunk and data-ink — removing decorative elements that add no information; increasing the ratio of meaningful ink to total ink (Tufte, VDQI, 1983).
Preattentive attributes — using color, size, position, and shape to direct attention before conscious processing (Knaflic, Storytelling with Data, 2015; Ware, Information Visualization, 2004).
Dashboard layout and KPI design — organizing multiple views for rapid comprehension; Few (Information Dashboard Design, 2006).
Truthful encoding — detecting and fixing charts that lie through truncated axes, cherry-picked ranges, dual axes, and misleading proportions (Cairo, How Charts Lie, 2019).
Chart accessibility — colorblind-safe palettes for data (distinct from brand palettes), alt-text for charts, pattern + color redundancy.
Marks and channels — the rigorous encoding framework: what data type maps to which visual channel (Munzner, Visualization Analysis & Design, 2014).

Not the brand or UI color palette (core color mode), overall page composition and visual hierarchy (core audit mode), or data-display tables as a UI interaction pattern (use usability).

Rules

Standing rules for every data visualization decision. Kept separate so they don't dissolve into the procedure.

data-viz

When this applies

Rules