← All projects Β· Retired

Writing Compare

Retired

Compared any two pieces by vocabulary, style, structure and theme.

Compare any two pieces in the writing collection across vocabulary, style, structure, and thematic dimensions. Also generates a full pairwise similarity matrix for collection-wide analysis.

Vocabulary

  • Jaccard similarity β€” overlap of unique word sets
  • Distinctive words β€” high-frequency words unique to each piece
  • Shared top vocabulary β€” most common shared content words

Style

  • Sentence length (mean and variance)
  • Punctuation density (em dashes, questions, colons, semicolons)
  • Style distance β€” normalized difference across features

Structure

  • Paragraph shape (ascending, descending, diamond, even)

Themes

  • Shared bigrams and trigrams (phrase-level overlap)
  • Composite similarity β€” weighted blend of all dimensions

Rendered from the project's README.md, with setup instructions and anything private left out.