SHARE Score
The 25 universal signals that measure how well research datasets are shared. Each dataset is scored 0-100 across 5 FAIR-aligned buckets.
SHARE Score Formula
SHARE = (signals present / 25) × 100
Fixed denominator of 25 — unsupported signals score zero. Each signal contributes equally. Total: 0-100.
Key Innovation: Repository-Adaptive Pledges
Each repository pledges the signals it supports. Scores are calculated as the fraction of pledged signals that contain valid data, multiplied by 100. This separates deposit-time signals (S, H, A, E) from outcome signals (R), so repositories can't game scores by only tracking usage metrics.
Signal Derivation Methodology
Every signal is grounded in established metadata standards. Each was derived by surveying four authoritative sources: DataCite Metadata Schema 4.6 (20 properties), Dublin Core (15 core elements), schema.org/Dataset (Google Dataset Search), and the RDA FAIR Data Maturity Model (41 indicators). A signal is included when it meets convergent support from 2+ independent standards, maps to a specific FAIR sub-principle, is measurable across diverse repositories, and has discriminative power. Click any signal below to see its full derivation and justification.
Note on empirical data: Per-signal prevalence statistics use the full Zenodo corpus (n=1.3M datasets) as the primary validation source because it is the largest general-purpose repository with structured DataCite metadata, natural variance across all 25 signals, and available citation/usage data. Citation Rate Ratios use the Zenodo 2016 deposit cohort (n=48,771) to ensure sufficient citation follow-up time (8+ years). The ratio is calculated as mean citations when a signal is present divided by mean citations when absent. Domain-specific repositories exhibit ceiling effects (e.g., SRA mean SHARE: 84.7) or restricted variance (e.g., Dryad CC0 mandate eliminates Access signal variation) that preclude per-signal discrimination. Cross-repository validation confirming framework generalizability is reported in the supplementary materials. The validation script is open-source: validate_v2.py.
25 Universal Signals
FAIR Principle Crosswalk
How each FAIR sub-principle maps to specific SHARE signals. System-level requirements (F3, F4, A2) are handled by repository pledges, not dataset-level scoring.
Scoring Tiers
Bucket-Level FAIR Alignment
Add Your Repository
Map your repository's metadata fields to the 25 universal signals and join the SHARE registry.