README.md

metadata

license: cc0-1.0
language:
  - multilingual
tags:
  - crossref
  - affiliations
data_files:
  - split: author
    path: data/author-*.parquet
  - split: editor
    path: data/editor-*.parquet
  - split: translator
    path: data/translator-*.parquet
  - split: chair
    path: data/chair-*.parquet
  - split: institution
    path: data/institution-*.parquet
  - split: project_investigator
    path: data/project_investigator-*.parquet
  - split: project_lead_investigator
    path: data/project_lead_investigator-*.parquet
  - split: project_co_lead_investigator
    path: data/project_co_lead_investigator-*.parquet
  - split: all
    path: data/all-*.parquet
  - split: normalized
    path:
      - data/normalized-00000-of-00034.parquet
      - data/normalized-00001-of-00034.parquet
      - data/normalized-00002-of-00034.parquet
      - data/normalized-00003-of-00034.parquet
      - data/normalized-00004-of-00034.parquet
      - data/normalized-00005-of-00034.parquet
      - data/normalized-00006-of-00034.parquet
      - data/normalized-00007-of-00034.parquet
      - data/normalized-00008-of-00034.parquet
      - data/normalized-00009-of-00034.parquet
      - data/normalized-00010-of-00034.parquet
      - data/normalized-00011-of-00034.parquet
      - data/normalized-00012-of-00034.parquet
      - data/normalized-00013-of-00034.parquet
      - data/normalized-00014-of-00034.parquet
      - data/normalized-00015-of-00034.parquet
      - data/normalized-00016-of-00034.parquet
      - data/normalized-00017-of-00034.parquet
      - data/normalized-00018-of-00034.parquet
      - data/normalized-00019-of-00034.parquet
      - data/normalized-00020-of-00034.parquet
      - data/normalized-00021-of-00034.parquet
      - data/normalized-00022-of-00034.parquet
      - data/normalized-00023-of-00034.parquet
      - data/normalized-00024-of-00034.parquet
      - data/normalized-00025-of-00034.parquet
      - data/normalized-00026-of-00034.parquet
      - data/normalized-00027-of-00034.parquet
      - data/normalized-00028-of-00034.parquet
      - data/normalized-00029-of-00034.parquet
      - data/normalized-00030-of-00034.parquet
      - data/normalized-00031-of-00034.parquet
      - data/normalized-00032-of-00034.parquet
      - data/normalized-00033-of-00034.parquet

Crossref unique affiliations

Exact organization strings extracted from Crossref snapshot 2026-07.

For the nine exact-string splits, no trimming, case folding, or Unicode normalization is performed. Repeated leaf occurrences are counted, and empty decoded strings alone are excluded.

Normalized split

The normalized split groups the exact strings in all after applying this normalization contract:

1. Transliterate Unicode text to ASCII with Unidecode.
2. Convert letters to lowercase.
3. Remove punctuation while preserving internal whitespace.
4. Trim surrounding whitespace.

Rows that become empty are excluded and counted in normalization_metadata.json. Each output row contains the normalized string, the contributing Crossref sources in registry order, the summed occurrence count, and the number of distinct raw affiliations that collapsed into it.