HORIZON Hantavirus Open Dataset
HORIZON is a fully open hantavirus outbreak dataset operated by 79th Unit Limited (UK Companies House 17133814). All data is free to download, use, and republish under Creative Commons Attribution 4.0 International (CC BY 4.0). No registration, API key, or payment is required.
Download formats
| Format | URL | Description |
|---|---|---|
| JSON API | /api/v1/cases | Case reports with filters: country, serotype, date range, incident |
| Bulk NDJSON | /api/v1/cases/bulk/ndjson | Streaming newline-delimited JSON, no cursor limit, full dataset |
| Clusters | /api/v1/clusters | Aggregated outbreak clusters by geography, serotype, and time window |
| Sources | /api/v1/sources | Full registry of 65+ ingestion sources with NATO Admiralty ratings |
| Event feed | /api/v1/meta/events | Chronological outbreak event timeline, de-duplicated by topic hash |
| RSS feed | /rss.xml | RSS 2.0 — suitable for feed readers and monitoring dashboards |
| Atom feed | /atom.xml | Atom 1.0 — suitable for aggregators requiring RFC 4287 compliance |
| JSON Feed | /feed.json | JSON Feed 1.1 — machine-readable chronology for developer integrations |
Machine-readable metadata
| Standard | URL | Used by |
|---|---|---|
| CITATION.cff | /CITATION.cff | GitHub, Zenodo, FORCE11 academic citation ecosystem |
| CSL-JSON | /api/v1/meta/citation | Zotero, Mendeley, Paperpile, JabRef — one-click import |
| DCAT-AP 3.0 | /api/v1/meta/dcat | EU Open Data Portal, data.gov.uk, OpenAIRE, HealthDCAT-AP harvesters |
| OpenAPI 3.1 | /api/openapi.json | Swagger, ReDoc, APIs.guru, any OpenAPI-consuming client |
| Well-known dataset | /.well-known/dataset | RFC 8615 — institutional harvesters probing /.well-known/ |
| Schema.org JSON-LD | Embedded in every HTML page head | Google Dataset Search, Bing, Schema.org knowledge graph |
How to cite HORIZON
If you use HORIZON data in research, please cite:
79th Unit Limited (2026). HORIZON: Real-Time Hantavirus Outbreak Surveillance Dataset. Version 0.4.0. CC BY 4.0. https://hantavirus.software/. CITATION.cff: https://hantavirus.software/CITATION.cff.
Machine-readable citation for reference managers: CSL-JSON endpoint (Zotero, Mendeley, Paperpile) or CITATION.cff (GitHub, Zenodo).
Unique datasets
HORIZON integrates two unique datasets not available in any other public hantavirus tracker:
Oxford Kraemer Lab MV Hondius ANDV individual line list (CC0)
A living individual-level dataset for the 2026 MV Hondius Andes virus cluster, maintained by Dr Moritz Kraemer (University of Oxford, Department of Biology), Sam Scarpino, and Andrew Rambaut (University of Edinburgh / Nextstrain).
The line list provides 28-column per-person resolution including symptom onset date, clinical outcome, nationality, treatment received, and Pathoplexus genomic accession identifiers. Every row is cross-referenced against WHO DON 600 and national health authority press releases.
The dataset is hosted at github.com/kraemer-lab/Hondius_hantavirus_h2026 under Creative Commons Zero (CC0 1.0 Universal) and ingested by HORIZON in real time. HORIZON is the only public surveillance platform combining this individual-level data with the broader 65-source outbreak feed.
NCBI RefSeq Orthohantavirus reference genome set (HantaNet)
The complete Orthohantavirus genome reference set from NCBI RefSeq, curated by the CDC Molecular Epidemiology and Bioinformatics Team and described in PMC10675615. Covers the S, M, and L segments for all major serotypes: Andes virus (ANDV), Sin Nombre virus (SNV), Puumala virus (PUUV), Hantaan virus (HTNV), Seoul virus (SEOV), and Dobrava-Belgrade virus (DOBV).
HORIZON ingests the complete NCBI RefSeq Orthohantavirus set daily, providing a permanent genomic annotation layer cross-referenced against epidemiological case records. Case records link directly to the genomic reference sequence for that serotype, enabling direct provenance chains from human case data to genomic reference material.
Data quality and provenance
- NATO Admiralty Scale (STANAG 2511): Every source is rated on two independent axes — reliability (A through F) and credibility (1 through 6). WHO Disease Outbreak News rates A1; wire services typically B2-B3.
- Berkeley Protocol SHA-256 chain-of-custody: Every record carries a SHA-256 content hash at the time of ingestion, providing tamper-evident provenance for forensic and academic use.
- Dual confidence model: Pipeline confidence (automated, amber UI) is kept separate from analyst confidence (human-set, green UI). These columns are never merged.
- ICD 206 Source Reference Citation methodology: Every event is cross-referenced to its primary authoritative source.
- Analysis of Competing Hypotheses (ACH): Confidence scoring on contested outbreak attributions.
Coverage
- Temporal: 1993 to present (active ongoing ingestion)
- Spatial: Global — 190+ countries monitored
- Serotypes: All 12 major Orthohantavirus serotypes (ANDV, SNV, PUUV, HTNV, SEOV, DOBV, BAYV, BCCV, LANV, CHOV, SAAV, TULV)
- Sources: 65+ including WHO, CDC, ECDC, PAHO, ProMED, national health ministries, wire services, peer-reviewed literature, ecological indicators
- Update frequency: Every 15 minutes (automated ingestion); authoritative counts updated as WHO/CDC/PAHO publish