data-vizlisted

# Data Viz (honest, web-native, Tufte-informed) Make web-native charts whose marks map honestly to data, fit the comparison task, and remain tokenized, responsive, and accessible. Use specialist Tufte tooling when present. ## The decision order 1. State the comparison question, source/grain, missingness, and uncertainty. 2. Prefer a table when exact lookup matters and the value set is small; prefer a chart when shape, change, or comparison is the task. 3. Use honest encoding (zero-baseline bars; 1-D length/position; lie factor ≈1), direct labels, and no non-data ink. 4. Provide tokenized responsive contrast-safe chart/table parity; interactive data is keyboard reachable. ## Data-quality gate: decide what the marks are allowed to claim Before choosing a genre, write down the comparison question, unit, time grain, source, and the status of every absent value. A missing month can mean **zero**, **not observed**, **suppressed**, **not applicable**, or **not yet reported**; those are different data states and must not share a mark, a line segment, or a table cell treatment. Also name what an uncertainty band means (for example, confidence interval, forecast range, or measurement error), the interval level if known, and whether all observations use the same method. If values, units, provenance, or interval meaning are absent, stop at a representation specification and data-request checklist. Do not invent a trend, interpolate a gap, or manufacture an impressive empty ch