S-index Real-World Testing and Validation Dataset

Overview

This dataset contains dataset metadata, FAIR scores, citations, mentions, and research field data collected/generated during our real world testing and validation of the S-index. The analysis of this dataset was included in our NIH S-index Challenge Phase 2 proposal. The code used to collect this data and additional details about the data collection methods and data structure are available in the related GitHub repository. We refer to the S-index Hub for more information about our S-index, the Challenge, and other related resources. 

Details

All data files are in NDJSON format. Data was collected between October 2025 and January 2026. A DuckDB file combining this data is available here.

Folder
Description

metadata
Contains reduced metadata of 49M+ datasets from the Datacite database and 51k+ datasets from the Electron Microscopy Data Bank (EMDB). Includes all datasets registered/published on or before September 30th, 2025. Data is in NDSON format following the DataCite Metadata Schema.

fair_scores
Contains FAIR scores of all the datasets calculated using F-UJI (between October 2025 and January 2026). 

citations
Contains 7.7M + unique citations to our datasets collected from the Make Data Count Data Citation Corpus, OpenAlex, and DataCite. 

mentions
Contains 91k+ mentions to our datasets collected from the README files of GitHub repositories using Software Data, Model Cards of AI/ML models on Hugging Face, and patents from the USPTO. 

research_fields
Contains research fields assigned to our datasets following the OpenAlex Topics taxonomy. Fields assigment comes from OpenAlex and a custom classifier we developed.