Contributing citations to the Data Citation Corpus
Data model
The data model for data citations in the Data Citation Corpus incorporates the properties below.
Dataset
| Element | Required? | Description |
| Identifier | Required | DOI, accession number, or URI for the dataset being cited |
| Repository | Required | The name of the repository, database or location on the internet where the dataset being cited is hosted |
| Affiliation | The organizational or institutional affiliation of the creator of the dataset that is cited
| |
| Funder | Funding information for cited dataset
| |
| Subjects | Subject information for cited dataset | |
| Title | Title of the cited dataset |
Article
| Element | Required? | Description |
| Identifier | Required | DOI for the journal article or preprint that includes the mention to the dataset |
| Publication date | The date at which the citation from the article to the dataset took place – For metadata purposes this is the article publication date | |
| Publisher | Publisher for the article that cites the dataset | |
| Journal | Journal or preprint server for the article that cites the dataset |
Data citation
| Element | Required? | Description |
| Provenance | Required | Organization or community that is the source for the data citation i.e. that identified and asserts the relationship between the dataset and the article |
Data ingestion
Contributors to the Data Citation Corpus are advised to provide the following information for the ingestion of their data citations into the Corpus:
| Item | Description |
| Scope | Data-article pairs, i.e. one object is a dataset, the other object is a journal article or a preprint. Data sources that include data as well as other resources should provide information on how to identify data records as part of their information. |
| Metadata elements | Required:
Optional: Additional metadata elements mapping to the data model above |
| Provenance | Confirmation of how the data source should be designated in the Corpus Description of methodology to assert the data-paper link, either by documentation provided directly or by published resources that outline the methodology |
| Format | Data citations in a CSV or JSON file One article-data pair per line (i.e. not one to many) |
Data checks
The following checks are completed on data citations prior to their inclusion in the Data Citation Corpus:
- Standardization of identifiers
- Standardization of repository names
- De-duplication of data-article pairs
- Identification of metadata errors e.g. publication date errors, verification that the dataset DOI and the article DOI are not the same
- Accession number format verification against repository identifier structure
