Synthetic COVID-19 booster effectiveness datasets for FAIR Data Train data visiting scenarios

Synthetic datasets accompanying the paper Extending the FAIR Data Train architecture through fixed and adaptive data visiting implementations. Each RDF file (N-Triples) is the local dataset of one simulated FAIR Data Station (a hospital, a registry or a regional authority) that trains visit in three use cases: UC1, fan-out and pool with capability matching: four hospital cohorts and a registry without vaccination data. UC2, a result-dependent sequence followed by fan-out: a Regional Hospital station, a Hospital Booster Registry with the identifier mappings, and four hospital cohorts. UC3, train-directed sequential visiting: four hospital cohorts with referral records. Each hospital cohort follows a test-negative design for COVID-19 booster vaccine effectiveness against hospitalisation (21,400 admitted persons across four sites, 2026 variant wave). Each site has a known true effectiveness, and ground-truth.json records the generation parameters and the results each use case must reproduce. All records are synthetic; no real person or hospital is described. The files are deterministic: the generators in the code record (https://doi.org/10.5281/zenodo.23915575) reproduce every file byte for byte. README.md contains the data dictionary.