---
license: cc0-1.0
task_categories:
  - text-classification
  - feature-extraction
language:
  - en
tags:
  - arxiv
  - crossref
  - citations
  - preprints
  - scholarly-communications
  - bibliometrics
  - doi
pretty_name: Crossref arXiv Citations
size_categories:
  - 100K<n<1M
configs:
  - config_name: all
    data_files:
      - split: train
        path: data/all.parquet
    default: true
  - config_name: asserted
    data_files:
      - split: train
        path: data/asserted.parquet
  - config_name: mined
    data_files:
      - split: train
        path: data/mined.parquet
---

# Crossref arXiv Citations

A dataset of arXiv preprints and their citations extracted from Crossref
metadata, validated against DataCite records.

## Dataset Description

This dataset maps arXiv works to the works in Crossref that cite them. Each
record represents an arXiv preprint with all known citations from
Crossref-registered works.

Built from the Crossref Metadata Plus monthly snapshot `2026-07` and the
DataCite monthly data file `2026-07` with
[crossref-citation-extraction](https://github.com/cometadata/crossref-arxiv-citation-extraction)
version 2.1.0.

### Dataset Configurations

| Config | Description | arXiv Works | Citing Works | Reference Matches |
|--------|-------------|-------------|--------------|-------------------|
| `all` | All validated citations (default) | 1,042,011 | 5,999,695 | 6,028,760 |
| `asserted` | Citations where the DOI was explicitly provided by the publisher or matched by Crossref | 160,089 | 452,501 | 454,186 |
| `mined` | Citations where the DOI was extracted from unstructured text | 998,112 | 5,548,061 | 5,574,574 |

### Loading the Dataset

```python
from datasets import load_dataset

dataset = load_dataset("cometadata/crossref-arxiv-citations")
asserted = load_dataset("cometadata/crossref-arxiv-citations", "asserted")
mined = load_dataset("cometadata/crossref-arxiv-citations", "mined")
```

## Data Schema

Each record contains:

| Field | Type | Description |
|-------|------|-------------|
| `arxiv_doi` | string | The DOI for the arXiv work (format: `10.48550/arXiv.{id}`) |
| `arxiv_id` | string | The arXiv identifier (e.g., `1412.6980`) |
| `reference_count` | integer | Total number of reference match instances |
| `citation_count` | integer | Number of unique citing works |
| `cited_by` | array | List of citing work objects |

Each citing work object in `cited_by` contains:

| Field | Type | Description |
|-------|------|-------------|
| `doi` | string | DOI of the citing work |
| `provenance` | string | Highest-confidence provenance among this work's matches (`publisher` > `crossref` > `mined`) |
| `matches` | array | Reference matches from this citing work |

Each match contains:

| Field | Type | Description |
|-------|------|-------------|
| `provenance` | string | How this match was obtained (see below) |
| `raw_match` | string | The exact substring that matched an arXiv identifier pattern |
| `reference` | struct | The Crossref reference metadata (string-valued fields; absent keys are null) |

### Provenance Values

- `publisher` - the reference carries an explicit arXiv DOI with `doi-asserted-by: publisher`
- `crossref` - the reference carries an explicit arXiv DOI with `doi-asserted-by: crossref`
- `mined` - the arXiv identifier was extracted from unstructured reference text

The `asserted` config contains only `publisher` and `crossref` matches; the
`mined` config contains only `mined` matches. A citing work appears in a
config only with its matches of that config's provenances, and all counts are
recomputed per config.

## Extraction Process

arXiv references are extracted from the Crossref snapshot by scanning
reference metadata for arXiv identifiers in modern (`arXiv:2403.03542`),
legacy (`arXiv:cs.DM/9910013`), DOI (`10.48550/arXiv.2403.03542`), and URL
(`arxiv.org/abs/2403.03542`) forms. Identifiers are normalized (lowercase,
version suffixes removed). Extracted arXiv DOIs are validated against an
index of DOIs built from the DataCite monthly data file `2026-07`;
only arXiv works registered in DataCite are included.

## Data Sources

- Crossref Metadata Plus monthly snapshot `2026-07`
- DataCite monthly data file `2026-07`
- Extraction tools: [cometadata/crossref-arxiv-citation-extraction](https://github.com/cometadata/crossref-arxiv-citation-extraction)

## Limitations

- Only includes citations registered in Crossref metadata
- Reference parsing depends on successful identifier extraction from
  unstructured references
- The `mined` subset may contain lower-quality matches from noisy reference
  text

## License

This dataset is released under
[CC0 1.0 Universal](https://creativecommons.org/publicdomain/zero/1.0/).

## Citation

```bibtex
@dataset{crossref_arxiv_citations,
  title = {Crossref arXiv Citations},
  author = {Cometadata},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/cometadata/crossref-arxiv-citations}
}
```