The hunt for research data: Development of an open-source workflow for tracking institutionally-affiliated research data publications

TL;DR

This paper presents an open-source workflow for tracking institutionally-affiliated research data publications across multiple platforms to improve data findability and support research data stewardship.

Contribution

It introduces a multi-API workflow for institutional data discovery, addressing challenges of inconsistent metadata and expanding the ability to find datasets regardless of DOI or affiliation metadata.

Findings

Retrieved over 4,000 datasets across 70 platforms

Identified major gaps due to inconsistent metadata practices

Demonstrated feasibility of multi-API institutional data tracking

Abstract

The ability to find data is central to the FAIR principles underlying research data stewardship. As with the ability to reuse data, efforts to ensure and enhance findability have historically focused on discoverability of data by other researchers, but there is a growing recognition of the importance of stewarding data in a fashion that makes them FAIR for a wide range of potential reusers and stakeholders. Research institutions are one such stakeholder and have a range of motivations for discovering data, specifically those affiliated with a focal institution, from facilitating compliance with funder provisions to gathering data to inform research data services. However, many research datasets and repositories are not optimized for institutional discovery (e.g., not recording or standardizing affiliation metadata), which creates downstream obstacles to workflows designed for theoretically comprehensive discovery and to metadata-conscious data generators. Here I describe an open-source workflow for institutional tracking of research datasets at the University of Texas at Austin. This workflow comprises a multi-faceted approach that utilizes multiple open application programming interfaces (APIs) in order to address some of the common challenges to institutional discovery, such as variation in whether affiliation metadata are recorded or made public, and if so, how metadata are standardized, structured, and recorded. It is presently able to retrieve more than 4,000 affiliated datasets across nearly 70 distinct platforms, including objects without DOIs and objects without affiliation metadata. However, there remain major gaps that stem from suboptimal practices of both researchers and data repositories, many of which were identified in previous studies and which persist despite significant investment in efforts to standardize and elevate the quality of datasets and their metadata.