Harborsignal

HarborSignal (harborsignal.ai) — AIS-gap dark-fleet event dataset: hourly polls of every oil tanker and LNG carrier broadcasting AIS inside monitored energy-port zones, differenced poll-to-poll so that a vessel present in one poll and absent in the next without a logged departure becomes a timestamped dark-vessel event. Each event carries MMSI and vessel identity, port zone, last-seen timestamp, last-known speed and movement state (at anchor / underway), AIS identity-quality grade, port sanctions-risk tier per zone, and a 0-10 composite sanctions-risk score built from port risk, last-contact …

Harborsignal offers (Alternative, Geospatial, Supply Chain, Sentiment) — The detection layer produces 24 polls per vessel-zone per day. With 3 vessels flagged in the current window across 4 monitored zones, event volume is modest — order tens of scored gap events per day, i.e. roughly 10^3-10^4 events per year. The underlying presence panel is far larger: hourly snapshots for every tanker/LNG carrier visible in four major zones implies roughly 10^5-10^6 vessel-presence observations per year — a few hundred MB of structured rows, not a big-data asset but a rare labelled one..

Sanctions-evasion and compliance screening for P&I clubs, shipowners, charterers, ports and traders needing to know whether a counterparty vessel has gone dark at a high-risk zone; secondary-sanctions risk models and designation-prediction (does a gap event precede an enforcement listing?); crude-origin-provenance models; physical-crate alpha where dark-fleet activity at Ras Tanura or Singapore front-runs observable supply; maritime insurance underwriting; ML training of dark-vessel detectors from AIS (the labelled event plus absent-negative pair is the scarce ingredient); port-congestion and bunkering-demand nowcasting.

The data is with 1 years of history.

Coverage spans US, UK; Oil & Gas Storage & Transportation; alternative, geospatial, supply_chain, sentiment; equities, commodities, options.