Open data

Baltic Sonar publishes its current state and a rolling 30-day incident window as documented, stable data — so researchers and journalists can cite it, verify it, and build on it. The schema is kept stable on purpose: dashboard changes will not break your pulls.

⚠ This is automated aggregation with AI-generated analysis, not verified intelligence. Severity, attribution and corroboration are automated estimates. Use as situational context, not as a sole basis for any decision.

Endpoints

/data/current.json
Current three indices, per-domain risk, and the source-class table. Updates ~every 30 min.

/data/events-30d.json
Every incident from the last 30 days, with category, tags, severity, attribution, evidence type, corroboration and the credibility classes that reported it. JSON.

/data/events-30d.csv
The same 30-day window as CSV, for spreadsheets and notebooks.

/data/events-30d.geojson
Map-ready GeoJSON with point geometry — drop into QGIS, Leaflet or a notebook.

Two honesty notes on the event window. Dates: every row carries date_basis, which says what its date is: published (the source's own publication date), measured (our own sensor reading), aggregator_seen (when an aggregator first observed the article — a third party's clock in 15-minute buckets, not a publication date), or collected (no date existed, so it is our collection time). Only published and measured belong in a time series; ts_estimated is true for the other two and is kept only for backwards compatibility. Coordinates: one incident can be reported by several outlets; every field of a record (lat, lon, title, severity) is taken from a single real report, never mixed across reports. Coordinates may still be a city or country centroid.

Licence & attribution

Released under CC BY 4.0. You may share and build upon it, including commercially, with attribution:

Data: Baltic Sonar (balticsonar.eu)

The pipeline code, model prompts and the full historical archive are not part of this release. The 30-day window is provided for research and citation; for a longer series, get in touch.

Vessel designation data shown on the site is derived from OpenSanctions (CC BY-NC 4.0) and is not part of this CC BY 4.0 release. It is published separately at /data-nc/vessel-designations.json under CC BY-NC.

Practically: the files above carry no vessel-designation field, so a commercial reuser of this CC BY release never inherits an NC obligation. If you need the designations, take them from the NC endpoint and honour its non-commercial terms and its own attribution.

Example record

{
 "id": "ais_11f844fbb79f5b488ddb",
 "date": "2026-10-17",
 "datetime": "2026-10-17T14:30:01.424648Z",
 "ts_estimated": false,
 "date_basis": "measured",
 "category": "NAVAL",
 "tags": [
  "subsea-cable",
  "anchoring",
  "naval-vessel"
 ],
 "country": "multi",
 "severity": 50,
 "lat": 59.992,
 "lon": 24.978,
 "attribution": "UNKNOWN",
 "evidence": "FACT",
 "corroboration": 1,
 "source_classes": [
  "A"
 ],
 "title": "Vessel loitering near critical undersea infrastructure (Finland Estonia Connection 2)",
 "source_url": null
}

Fields are documented inline in each JSON file under schema. Generated 2026-10-17 14:30 UTC.