The Ocean Dataset — SeaSignals

Layer 01 · Vessels

Vessel AIS tracking

Positions, identities and behaviour for ships at sea. We merge land-based and satellite AIS, remove duplicates and errors, and add derived voyages and behaviour flags.

LatencySeconds to minutes

HistoryMulti-year archive

KeyMMSI · IMO

Feature	What you get
`positions`	Live lat/lon, speed, course, heading, navigational status
`tracks`	Historical tracks, cleaned of noise, for replay and analytics
`profiles`	Name, type, flag, dimensions, call sign
`voyages`	Detected departures and arrivals, port calls, predicted ETA
`events`	AIS gaps ("going dark"), encounters, loitering, likely spoofing, fishing activity

Layer 02 · Storms

Storm tracking

Tropical cyclones and severe marine weather, combined from official agency advisories and turned into geometry you can query: tracks, forecast cones and wind radii.

LatencyPer advisory cycle

HistoryDecades of best tracks

KeyStorm ID

Feature	What you get
`active`	Current systems with position, pressure, max winds, category
`forecast`	Forecast track, uncertainty cone, wind radii as GeoJSON
`marine_wx`	Gale and storm warnings, pressure and wind fields
`archive`	Historical best tracks for replay and climatology
`exposure`	Vessels, ports and zones inside a storm's path (cross-layer)

Layer 03 · Waves & ocean state

Wave & ocean conditions

Gridded wave and ocean model output, checked against real buoy readings. Get the sea state for any point, route or area, now or forecast days ahead.

LatencyHourly / model runs

ForecastUp to ~10 days

KeyGrid cell · buoy ID

Feature	What you get
`waves`	Significant wave height, peak period, direction, swell partitions
`wind`	10 m wind speed and direction, gusts
`ocean`	Sea surface temperature, currents, tides, sea ice
`buoys`	Live observations from moored and drifting buoys
`point / route`	Forecasts for a point or along a route, ready to use

Layer 04 · Ports

Port signals

Port activity derived from vessel behaviour around berths and anchorages. Shows congestion and throughput before official statistics are published.

LatencyNear real time

CoverageMajor global ports

KeyUN/LOCODE

Feature	What you get
`congestion`	Vessels at anchor vs. at berth, queue length, waiting time
`calls`	Arrivals, departures, dwell and turnaround times
`expected`	Predicted arrivals from inbound vessel ETAs
`throughput`	Call-based activity indices by vessel class (tanker, bulk, container)
`anomalies`	Week-over-week spikes and drops, weather-related disruptions

Layer 05 · Marine life

Marine life tracking

Movement and sighting data for whales, sharks, turtles and seabirds, aggregated from research and biodiversity programmes and shared responsibly.

LatencyDelayed by design

HistoryMulti-year tracks

KeyTaxon · tag ID

Feature	What you get
`tracks`	Tagged-animal movement, with time delay and location blurring where needed
`sightings`	Occurrence records and density heatmaps by species
`corridors`	Seasonal migration corridors and feeding grounds
`strike_risk`	Where vessel traffic overlaps whale presence (cross-layer)
`zones`	Marine protected areas and seasonal management areas

The shared spine

What makes it one dataset

Every layer is built on the same four foundations. That's why cross-layer joins are a single query parameter instead of a data engineering project.

Space: one global grid

Observations are indexed on a hierarchical hexagonal grid (H3) as well as raw coordinates. Zoom from ocean basin down to harbour scale, and aggregate any layer to any cell size.

Time: one UTC timeline

Each record has both an observation time and an ingestion time, so you can query what was true, or what was known, at any moment. That matters for backtests and replays.

Entities: shared IDs

Vessels, ports, storms, buoys, zones and species are entities with stable IDs. A port call links a vessel to a port, and an exposure links a vessel to a storm.

Provenance: always attached

Source, licence, latency, processing version and a confidence score travel with every value. Filter by source, or exclude anything below a confidence level.

Sources

Built on trusted ocean data

We combine open scientific and government data with licensed commercial feeds where needed. Licences and attribution are tracked for every source.

Layer	Planned source types
Vessels	Land-based AIS networks, satellite AIS providers
Storms	NOAA NHC, JTWC and regional meteorological agencies
Waves	NOAA WAVEWATCH III, Copernicus Marine, ECMWF open data, NDBC and partner buoys
Ports	Derived from AIS with port and anchorage geofences, UN/LOCODE
Marine life	OBIS, Movebank, GBIF and research partners

Delivery

Use the data however you work

REST & GeoJSON API

Query by area, time, entity or grid cell. Join layers with one parameter.

Live streams

WebSocket and SSE feeds for positions, storm advisories and port events.

Alerts & webhooks

Geofences, storm proximity, wave thresholds and congestion spikes.

Bulk history

Partitioned Parquet and CSV exports, delivered to your cloud bucket.

Warehouse shares

Query directly from Snowflake, BigQuery or DuckDB.

Dashboards

Ready-made views for ports, storms, fleets and wildlife, plus your own.

Responsible data

Some signals need care.

Precise, real-time locations of endangered species can be misused. Sensitive wildlife data is delayed, blurred or limited to approved research partners, following each data provider's terms. SeaSignals is an analytics platform and is not for navigation. Always rely on official charts, forecasts and notices to mariners.

Ready to query the ocean?

Request early access to the API and sample datasets.

Start free trial See the API

One model ofthe entire ocean.