xd Crossword Corpus
xd.saul.pw (Saul Pwanson)
Over 6 million crossword clue-answer pairs, from the project that caught a crossword plagiarist.
A crossword research corpus in a simple text format, with puzzle metadata and similarity scores across tens of thousands of puzzles. The public downloads include every clue-answer usage grouped by publication year, plus 5,688 full puzzles published before 1965. It is the data behind the 2016 exposure of a puzzle editor who reused other constructors' grids.
Access
- Cost
- FreeFree zip downloads; source code on GitHub.
- License
- License not stated
- Sign-up
- No sign-up
- How you get it
- Bulk download.xd text, TSV
What's in it
- One row is
- One row per clue-answer usage; one file per pre-1965 puzzle
- Covers
- Mostly U.S. newspaper crosswords
- Years
- 1924-present for clues; full puzzles 1924-1964
- Updated
- Irregular
- Size
- Over 6,000,000 clue usages (67MB zip); 5,688 public puzzles
Quality
Unique and well structured, but a one-person project whose full puzzles are limited to the public domain.
- Modern puzzles are available only as clues, not grids
- Coverage of publications is uneven
Connects to
- Chronicling America: Early newspaper crosswords in their original pages
- Unicode Emoji Frequency: Another record of shifting everyday vocabulary
Questions it could answer
- Which answers (like ERIE or OREO) are most overused, and has that changed?
- How did clue style shift from definitions to wordplay over the decades?