Manually curated links between data papers and research papers

The csv file contains a manually curated dataset linking datasets, data papers and research papers. The point of departure for manual linking was the data paper which describes a dataset. We then carried out a step-by-step research on various research platforms (such as Google Scholar or Dimensions) to identify the research paper which can be considered to be an output of <i>the same research effort</i>.The manual curation process was developed from the heuristic rules established in McGillivray et al 2020. We linked a research paper to a data paper if at least one of the following four conditions was satisfied: (1) the research paper appeared in the reference list of the data paper; (2) the research paper was cited in the dataset repository; (3) the research paper cited the data paper; (4) the research paper cited the dataset. Additionally, the following two conditions also had to be satisfied: (a) at least one person was an author of both the data and the research paper; and (b) the research paper was a substantial, analytical interpretation of the dataset associated with the data paper.Please see a discussion of the heuristic rules and the automation of the linking process in our paper "Finding the Paper Behind the Data: Automatic Identification of Research" accepted for presentation at the Workshop for Artificial Intelligence for Scientific Publications (WASP) 2025 associated with the International Joint Conference on Natural Language Processing &amp; Asia-Pacific Chapter of the Association for Computational Linguistics (IJCNLP-AACL).Articles related to Data Publications<br>