Exploring integration of Scholar Data dataset metrics (D-Index, citations, FAIR score)

[megasanjay]

Hi PhysioNet team!

We’re working on a platform called Scholar Data as part of the NIH S-Index Challenge. One component of our work is a computed dataset-level impact index (D-Index), alongside citation counts, FAIR/FUJI scores, and mention metrics.
We’ve computed these metrics for all publicly identifiable datasets up to September 2025, including PhysioNet datasets with DOIs.
We wanted to ask whether you might be interested in optionally integrating these metrics into PhysioNet dataset pages. The goal would be to provide users with an at-a-glance, standardized signal of dataset impact and reuse, without changing PhysioNet’s existing curation model.

For any dataset with a DOI, we can provide:

• Dataset Index (D-Index) score
• Total citation count
• Total mentions (e.g., scholarly references, GitHub repositories, and other non-academic sources)
• FAIR/FUJI score with versioning and evaluation date

Integration options

We’re intentionally keeping integration lightweight. Two options that may be useful:

For concreteness, the examples below use a DOI from the PhysioNet MIMIC-IV dataset, which we’ve included in our September 2025 snapshot.

1. Embed (iframe)

A small embeddable widget that renders the dataset metrics:

<iframe
  src="https://beta.scholardata.io/embed/d-index?doi=10.13026/kpb9-mt58"
  width="245"
  height="200"
  loading="lazy"
  title="Dataset Index embed"
></iframe>

This requires no backend changes and can be fully sandboxed.

FYI we also noticed that there is a metrics panel that was added by #2556. If interested, we can figure out a way to stylize the embed with something that looks more native to your platform.

2. JSON API endpoint

If you prefer full control over presentation, we expose a DOI-based JSON endpoint:

GET https://beta.scholardata.io/api/v1/datasets/by-doi?doi=10.13026/kpb9-mt58

Example response:

{
  "datasetId": 19256468,
  "totalCitations": 51,
  "totalMentions": 0,
  "fujiScore": {
    "score": 73.08,
    "evaluationDate": "2026-01-08T11:46:27.000Z",
    "metricVersion": "0.8",
    "softwareVersion": "3.5.1"
  },
  "latestDIndex": {
    "score": 4.095372884942452,
    "created": "2025-12-08T00:00:00.000Z"
  }
}

Notes

• All scores are computed, versioned, and time-stamped
• No tracking, ads, or user data collection
• This is exploratory; we’re mainly looking for feedback on whether this would be useful or appropriate for your users

If this sounds potentially interesting, we’d be glad to share more details on methodology, validation, or run a small pilot on a subset of datasets.

Thanks for your time, and for the work you do maintaining PhysioNet!

[tompollard]

Thanks @megasanjay, I'd love to help you to pilot this. We'll have a chat about the implementation and will follow up with you soon. Copying @Chrystinne who I hope can help out.

[Chrystinne]

Very interesting initiative, @megasanjay! And yes, @tompollard I'd be interested in exploring how we can best integrate these metrics into PhysioNet.