An OpenRefine extension for spotting data-quality problems in text — especially messy, multilingual, transliterated metadata.
Status: 41 text quality checks and linguistic analysis tools across encoding, scripts, normalization, diacritics, casing, punctuation, content, glossary, and index.
loupe will let you point OpenRefine at a column and get a plain-language report of the likely problems in it, so you can find and fix messy data without reading every cell by hand. It's aimed at multilingual and transliterated material — names and titles in Arabic, Persian, mixed scripts, and the like — but works on any text.
It will flag things such as:
- garbled or mis-encoded characters
- mixed or unexpected writing systems
- stray, hidden, or inconsistent spacing
- inconsistent accents and apostrophes in transliteration
- empty, placeholder, or obviously broken values
For end users: Download loupe.zip from releases, unzip, and drag the loupe folder into your OpenRefine extensions directory. See OpenRefine's extension installation docs for details.
- Clone and build:
make extension - Install:
make install - Restart OpenRefine
- Column dropdown → Facet → Loupe submenus
Requires: Leiningen, OpenRefine 3.10.x, macOS (default extensions path).
Click any column header → Facet. Five Loupe submenu groups provide text quality checks and linguistic analysis:
- Loupe: Quality — composite checks, encoding issues, whitespace analysis, normalization status, reconciliation readiness
- Loupe: Scripts & Characters — script detection, directionality, diacritics analysis, case patterns
- Loupe: Content Analysis — punctuation, quotes, numeric systems, emoji, word stats, ligatures, character categories
- Loupe: Glossary & Index — word index by prefix, word/bigram/trigram glossary
- Loupe: Reports — KWIC concordance, multi-column field report
Each check creates a text facet grouping cells by detected pattern or issue type. The concordance feature opens a dialog for keyword-in-context (KWIC) search across visible cells. The field report runs multiple checks across selected columns and displays frequency results in a tabbed overview.
See OpenRefine Data Quality Facets.md for the current list of functions.
CC-BY