Dataset Citation Finder

Orientation

BLE Citation Finder is an R workflow for identifying publications that cite, reference, or are associated with Beaufort Lagoon Ecosystems Long Term Ecological Research (BLE LTER) datasets.

The workflow builds a registry of BLE datasets from the Environmental Data Initiative (EDI) and BLE-funded datasets documented in Zotero. It then searches multiple scholarly sources for related publications.

Search results are compared with the BLE Zotero bibliography to identify potentially missing publications and dataset citation information.

Important: The workflow is intended as a citation discovery and review tool. Results should be manually verified before updating Zotero or EDI.

Features

The workflow:

• Builds a registry of BLE datasets from EDI and Zotero.
• Searches DataCite for dataset citation relationships.
• Searches OpenAlex for citations and dataset DOI references.
• Uses Crossref to retrieve missing publication metadata.
• Searches available publication PDFs for BLE keywords, dataset DOIs, and dataset URLs.
• Links publications to individual BLE datasets.
• Combines results found through multiple search methods.
• Compares results with the BLE Zotero bibliography.
• Identifies potentially missing publication-dataset relationships.
• Produces an Excel workbook for manual review.

API Keys

API credentials should be stored as environment variables and must not be committed to GitHub.

For example, the workflow uses the following environment variables:

• EDI_API_KEY
• OPENALEX_API_KEY

Usage

Run the main citation-finder R script.

The workflow first retrieves BLE datasets from the EDI scope knb-lter-ble and adds BLE-funded datasets archived elsewhere and documented in Zotero.

It then searches DataCite and OpenAlex for publications associated with each dataset. Crossref is used to complete missing publication metadata.

When publication PDFs are available, the workflow searches the full text for:

• BLE keywords
• Dataset DOIs
• Dataset URLs

The discovered publication-dataset relationships are then compared with Zotero, and an Excel workbook is created for review.

A publication may be associated with multiple BLE datasets. Therefore, comparisons are made using the combination of Paper DOI + Dataset DOI, rather than only the publication DOI.

Output

The workflow produces an Excel workbook containing the following tabs:

• BLE_Data_Registry
  Contains the BLE datasets included in the search.

• Publication_Search
  Contains publication-dataset relationships identified through DataCite, OpenAlex, and other searches.

• PDF_Results
  Contains BLE keywords, dataset DOIs, and dataset URLs found in available publication full text.

• Zotero_For_Website
  Contains existing publications and BLE dataset DOI information retrieved from Zotero.

• Missing_From_Zotero
  Contains publication-dataset relationships found by the search that may not yet be represented in Zotero.

• Zotero_Missing_Data_DOI
  Contains publications already in Zotero that may be missing an associated BLE dataset DOI.

• Zotero_Not_In_Search
  Contains known Zotero publication-dataset relationships that were not recovered by the automated search.

Interpreting Results

Different search methods provide different levels of evidence.

For example, a DataCite Citation relationship means that DataCite contains a structured relationship between the publication and dataset. It does not necessarily mean that the dataset DOI appears directly in the publication PDF.

Finding an exact dataset DOI or dataset URL in the publication full text provides stronger evidence of a dataset reference.

BLE keyword matches are useful for discovering candidate publications but do not by themselves confirm that a BLE dataset was used.

For this reason, records in Missing_From_Zotero should be manually reviewed before being added to Zotero.

Zotero Review

For each candidate in Missing_From_Zotero:

1. Open the publication using the paper DOI.
2. Check for the associated BLE dataset DOI or dataset information.
3. Review the Data Availability, Methods, References, and Acknowledgements sections when appropriate.
4. Review the Search_Source, Search_Method, and PDF search results.
5. If the publication-dataset relationship is confirmed, add or update the publication in Zotero.
6. Add the associated BLE dataset DOI to the Zotero Extra field when appropriate.

The script does not automatically modify Zotero or EDI records.

Limitations

The workflow may not find every publication using BLE data. Publications may omit dataset DOIs, PDFs may not be openly accessible, and external citation metadata may be incomplete.

For this reason, the workflow combines multiple search methods and uses manual verification as the final step.

Contributing

When updating the workflow:

• Update the date at the top of the README when appropriate.
• Update the documentation if search methods or outputs change.
• Do not commit API keys or credentials.
• Test the workflow before merging changes into the main branch.

Adapting the Workflow

The workflow was developed for BLE LTER but may be adapted by other LTER sites by changing:

• The EDI scope
• Zotero information
• Site-specific search terms