The composition of MW1899's 286,561 records by article type, as a squarified treemap.
Trust Block
- Evidence:
src/data/article-type-counts.json,src/data/article-type-examples.json, and MW source links. - Limitations: article types are overlapping classifier memberships, not an exclusive partition of MW records.
- Validation: generated by
npm run build-mw-depth; checked bynpm run validate-mw-depth,npm test, andnpm run build. - Owner repo:
csl-atlas. - Next use: use the chart to choose a source-linked MW sample or a follow-on comparison.
- Largest tile: compound sub-entries (126,360, 44.1%) — half the dictionary is compound members.
- Next: derived forms (72,119, 25.2%) — suffixed derivatives.
- Then Vedic-accented (47,598, 16.6%) — entries with the
/udātta marker; overlaps with other types. - Lexicographer-only (38,414, 13.4%) — entries whose only
<ls>citation isL.; also overlapping. - Smaller tiles: masculine nouns (19,204), other types.
Note: the types overlap (an entry can be both noun-m AND vedic-accented). The sizes are membership counts, not exclusive partitions. The total is not additive.
Each group lists three MW source records matching the same classifier used for the treemap counts. Links open the csl-orig MW source line for that record.
Source: CDSL mw.txt 2026-05-23. Static SVG: treemap-en.svg. CC-BY-SA-4.0.