FAIRagro RDI FAIRness Inventory

A web application and JSON API that publish FAIRness assessments of research data infrastructures (RDIs) in agrosystem research. Each repository is assessed against 20 criteria derived from the RDA FAIR Data Maturity Model, five per FAIR pillar (Findability, Accessibility, Interoperability, Reusability). Results come from structured interviews with repository managers, complemented by metadata harvested from re3data.org.

Scores are reported per pillar (e.g. "Accessibility 4/5"), never as a single composite FAIR index — see the methodology paper below.

Publication

Haleem A., Arend D., Etukala J.R., Rey Mazón E., Schmidt M., Jung J., Martini D., Usadel B., Neidiger C., Ulrich R., Lange M. (2026). A Standardized Methodology for FAIRness Assessment and Multi-Dimensional Scoring in Agrosystem Research Data Infrastructures. bioRxiv. https://doi.org/10.64898/2026.07.17.738915

If you use the data or the API, please cite the paper.

What you get

Route	Purpose
/	Searchable, filterable inventory; every card shows the 20-criterion diagnostic grid (green = met, red = not met, grey = not assessed)
/resource/repository/{RFId}	Repository profile: metadata, re3data record, per-pillar ratings, full criteria table
/resource/repository/{RFId}/fairness	Pillar bar chart + detailed assessment
/interview, /indicator/{1–20}	The questionnaire and the definition of each criterion
/downloads	Interview form, RDA specification, the raw interview spreadsheets
/schema	The MDS FAIRagro RDI metadata schema, browsable
/api-docs	API documentation with a live "try it" console

API

Read-only, no key, CORS enabled, described by OpenAPI 3.1 at /api/openapi. Highlights:

GET /api/resource/repository?q=soil&collection=Agricultural%20Sciences&sort=title&page=1&limit=20
GET /api/resource/repository/RFId001202604272            # profile, criteria, pillar scores, re3data
GET /api/resource/repository/RFId001202604272/fairness   # JSON-LD (schema.org DataCatalog)
GET /api/fairness            # matrix: every repository × 20 criteria
GET /api/fairness/csv        # same as CSV
GET /api/fairness/stats      # per-criterion compliance rates across the landscape
GET /api/indicators          # the 20 criteria with RDA codes
GET /api/facets              # subjects / collections with counts

Full reference: open /api-docs in the running app.

Data

Repository records are not stored in this repository. They ship in the npm package rf-rdis (one JSON file per RDI, Dataverse-style metadata blocks) and are bundled at build time. To publish updated assessments, release a new rf-rdis version, then rebuild and redeploy this app. The only data files kept here are in public/: the metadata schema, the indicator definitions and the downloadable interview documents.

Tech stack

Next.js 16 (App Router, route handlers for the API) · React 19 · MUI 9 · Recharts · TypeScript. Almost everything is pre-rendered at build time; only the search endpoint (/api/resource/repository?…) runs on the server because it reads query parameters.

Running it

Requires Node.js ≥ 20.9 (Node 22 LTS recommended) and npm.

Local development

npm ci
npm run dev          # http://localhost:3000

Production on a VM

git clone <this repository> repofinder && cd repofinder
npm ci
npm run build
PORT=3000 npm run start

Put a reverse proxy (nginx, Caddy, Traefik) in front of port 3000 for TLS, and keep the process alive with systemd or pm2, e.g. pm2 start npm --name rdi-inventory -- run start.

Set NEXT_PUBLIC_SITE_URL at build time to the public origin (default https://rdi-fairness-interviews.vercel.app); it is baked into the absolute links and JSON-LD @ids the API returns:

NEXT_PUBLIC_SITE_URL=https://inventory.example.org npm run build

Docker

docker build -t rdi-inventory --build-arg NEXT_PUBLIC_SITE_URL=https://inventory.example.org .
docker run -d --name rdi-inventory -p 3000:3000 rdi-inventory

The image builds the app in one stage and copies only the production output into a small node:22-alpine runtime that runs as an unprivileged user. Or with Compose:

services:
  rdi-inventory:
    build:
      context: .
      args:
        NEXT_PUBLIC_SITE_URL: https://inventory.example.org
    ports:
      - "3000:3000"
    restart: unless-stopped

Updating the data

npm install rf-rdis@latest
npm run build && npm run start      # or rebuild the Docker image

Project layout

app/
  page.tsx                  inventory (client component: search, filters, cards)
  resource/repository/      per-repository pages (SSG via generateStaticParams)
  indicator/, interview/    criteria definitions
  api/                      JSON API route handlers
    _lib/                   serializers, scoring, OpenAPI spec, response helpers
  api-docs/                 API documentation page + playground
  components/               UI components
    utils/                  data extraction, FAIRness scoring, colour/indicator constants
  theme.ts                  MUI theme (brand colours, typography)
public/
  indicators.json           the 20 criteria (single source of truth for the UI and API)
  MDS_FAIRagro_RDI_schema.json
  interviews/               completed interview spreadsheets

Accessibility note

Status colours use the Okabe–Ito green/vermillion pair, which remains distinguishable for the common forms of colour-vision deficiency, and every coloured icon also carries its status as text (tooltip and aria-label).

License

Application code: MIT. Assessment data: CC0 1.0 (see the license field of each record).