KennyChowww/openalex-snapshot-2026-05

OpenAlex Snapshot 2026-05 Processed Works and Citation Edges

This dataset repository contains a processed version of the OpenAlex Works snapshot from May 2026, together with derived citation-edge tables and selected downstream bibliometric datasets.

The repository is intended for large-scale scholarly metadata analysis, citation-network analysis, and bibliometric research. The main dataset displayed in the Hugging Face Dataset Viewer is light_openalex_works, which contains a lightweight version of OpenAlex Works split into Parquet shards.

Repository Contents

The repository is organized into several data subsets:

Subset	Description
light_openalex_works	Lightweight OpenAlex Works records. Default subset shown in the Dataset Viewer. Files are in light_OpenAlex_works/.
citation_edges	Expanded citation edges from OpenAlex referenced works. Files are in citation_edges/.
filtered_type_article_preprint_books	Filtered OpenAlex Works records for selected scholarly work types.
ai_article_preprint_with_institution_sector	AI-domain article/preprint sample enriched with institution-sector information.
final_ai_di5	Final AI-domain sample with DI5 values and related bibliometric fields.

Folder Structure

.
├── light_OpenAlex_works/
│   ├── works_light_part_00.parquet
│   ├── works_light_part_01.parquet
│   ├── ...
│   └── works_light_part_19.parquet
│
├── citation_edges/
│   ├── referenced_work_part_0.parquet
│   ├── referenced_work_part_1.parquet
│   ├── ...
│   └── referenced_work_part_19.parquet
│
├── filtered_type_article_preprint_books/
│   ├── part=0__data_0__filtered.parquet
│   ├── part=1__data_0__filtered.parquet
│   ├── ...
│   └── part=19__data_0__filtered.parquet
│
├── samples/
│   └── ai_article_preprint_with_institution_sector/
│       ├── openalex_ai_article_preprint_part-00000.parquet
│       ├── openalex_ai_article_preprint_part-00001.parquet
│       ├── ...
│       └── openalex_ai_article_preprint_part-00019.parquet
│
└── derived/
    └── di5_cd5_ai_article_preprint_with_institution_sector_v2/
        └── final_v2/
            ├── ai_article_preprint_with_di5_ref5_c5_pubyear_ge2010_education_company.parquet
            └── manifest_final.json

Main Dataset: light_openalex_works

The default dataset subset is light_openalex_works, stored in:

light_OpenAlex_works/

This subset contains lightweight OpenAlex Works records in Parquet format. It is designed as the main browsable dataset for the Hugging Face Dataset Viewer.

Main columns include:

Column	Description
work_id	OpenAlex Work ID.
publication_year	Publication year of the work.
publication_date	Publication date of the work, when available.
type	OpenAlex work type, such as article, preprint, review, or book.
topics	JSON-like OpenAlex topic metadata.
authorships	JSON-like OpenAlex authorship and institution metadata.
updated_date	OpenAlex record update date used in the processed snapshot.

Citation Edges

Citation edges are stored in:

citation_edges/

Each row represents one citation relationship between a focal work and one referenced work. The citation-edge table converts each work’s OpenAlex reference list into focal-to-referenced work pairs with focal publication dates.

Core columns include:

Column	Description
citing_work_id	OpenAlex Work ID of the focal/citing work.
citing_publication_year	Publication year of the focal/citing work.
citing_publication_date	Publication date of the focal/citing work.
referenced_work_id	OpenAlex Work ID of the referenced/cited work.

The edge list is filtered against the processed Works universe, so out-of-universe referenced works are removed. This helps prevent dangling cited-work nodes when constructing citation networks or calculating network-based indicators.

AI Article/Preprint Sample

The AI-domain sample is stored in:

samples/ai_article_preprint_with_institution_sector/

This subset contains AI-domain focal works with OpenAlex work type article or preprint. It also includes institution-sector information derived from authorship metadata.

Final DI5 Dataset

The final DI5 dataset is stored in:

derived/di5_cd5_ai_article_preprint_with_institution_sector_v2/final_v2/

DI₅ is a five-year disruption index calculated as:

DI₅ = (nᵢ − nⱼ) / (nᵢ + nⱼ + nₖ)

where, within a five-year citation window:

Term	Meaning
nᵢ	Number of later works that cite the focal work but do not cite its references.
nⱼ	Number of later works that cite both the focal work and its references.
nₖ	Number of later works that cite the focal work’s references but do not cite the focal work.

The final DI5 subset is intended for bibliometric analysis of AI-domain scholarly works.

Loading the Dataset

Load the default lightweight OpenAlex Works subset

from datasets import load_dataset

ds = load_dataset(
    "KennyChowww/openalex-snapshot-2026-05",
    "light_openalex_works",
    split="train",
    streaming=True,
)

for row in ds.take(3):
    print(row)

Load the citation-edge subset

from datasets import load_dataset

edges = load_dataset(
    "KennyChowww/openalex-snapshot-2026-05",
    "citation_edges",
    split="train",
    streaming=True,
)

for row in edges.take(3):
    print(row)

Query with DuckDB

import duckdb

repo = "hf://datasets/KennyChowww/openalex-snapshot-2026-05"

query = f"""
SELECT
  work_id,
  publication_year,
  publication_date,
  type
FROM '{repo}/light_OpenAlex_works/*.parquet'
LIMIT 10
"""

df = duckdb.sql(query).df()
print(df)

Intended Uses

This dataset may be useful for:

• large-scale bibliometric research;
• OpenAlex Works metadata analysis;
• citation-network construction;
• citation-window analysis;
• disruption-index research;
• AI-domain scholarly publication analysis;
• institutional and authorship metadata analysis.

Limitations

Users should consider the following limitations:

• The dataset inherits metadata coverage, disambiguation, and citation-link limitations from OpenAlex.
• topics and authorships are JSON-like columns and may require additional parsing.
• Citation-edge coverage depends on OpenAlex referenced-work metadata.
• The AI-domain sample depends on OpenAlex topic metadata and the selection criteria used to define AI-related works.
• DI5 values depend on citation-window definitions and available citation metadata.
• The derived subsets should not be treated as a complete representation of all scholarly works or all AI publications.

Dataset Maintainer

Maintainer:

Kenny CHOW

Citation

If you use this dataset, please cite this Hugging Face dataset repository and OpenAlex.

Suggested dataset citation:

@dataset{chow_openalex_snapshot_2026_05,
  title = {OpenAlex Snapshot 2026-05 Processed Works and Citation Edges},
  author = {Kenny CHOW},
  year = {2026},
  publisher = {Hugging Face},
  doi = {10.57967/hf/8850},
  url = {https://huggingface.co/datasets/KennyChowww/openalex-snapshot-2026-05}
}

Please also cite OpenAlex according to your publication venue's requirements.

License

This dataset is released under the Creative Commons Zero v1.0 Universal license.

This means the dataset may be used, copied, modified, and redistributed for both non-commercial and commercial purposes without asking permission. Citation is not legally required under CC0

license: cc0-1.0