AIS Data

AIS operates on two dedicated VHF frequencies (161.975 MHz and 162.025 MHz) using a Self-Organized Time Division Multiple Access (SOTDMA) protocol. This architecture allows vessels to autonomously reserve transmission slots, preventing signal collision in high-traffic zones. The system has an effective terrestrial range of approximately 20-30 nautical miles, though satellite-based AIS (S-AIS) receivers now provide global coverage by capturing signals from low-earth orbit.

AIS transmissions are bifurcated into two distinct message types with different refresh rates:

• Static Data (every 6 min): Maritime Mobile Service Identity (MMSI), IMO number, vessel name, call sign, dimensions, and ship type.
• Dynamic Data (2-10 sec): Position (lat/lon), Speed Over Ground (SOG), Course Over Ground (COG), rate of turn, and navigational status.
• Voyage Data (every 6 min): Destination port, estimated time of arrival (ETA), and draught. This temporal heterogeneity requires careful alignment when engineering features.

The Maritime Mobile Service Identity (MMSI) is a nine-digit numerical code that uniquely identifies each AIS transponder. It is the primary key for joining AIS data with other maritime datasets. However, MMSI spoofing and reuse are known issues. Engineering robust entity resolution pipelines requires cross-referencing MMSI with IMO numbers (a permanent, seven-digit hull identifier) and call signs to detect anomalies and maintain data integrity.

Raw AIS feeds are noisy. Common failure modes include:

• Gaps in transmission: Caused by VHF signal blockage, receiver saturation, or vessels disabling transponders (a known indicator of illicit activity).
• Positional jitter: GPS errors or deliberate manipulation.
• Implausible kinematics: Sudden jumps in position or impossible speed/heading combinations. Pre-processing pipelines must implement Kalman filters or heuristic rule sets to flag and interpolate anomalous records before they corrupt downstream trading signals.

AIS does not explicitly broadcast cargo contents. Instead, cargo type is inferred through a combination of:

• Vessel type code (e.g., crude oil tanker, bulk carrier, LNG tanker).
• Draught data: The distance between the waterline and the keel, which correlates with cargo weight. A laden tanker sits significantly deeper in the water than a ballasted one.
• Geospatial context: A vessel loitering near a known offshore lightering zone or making a port call at a specific terminal provides strong cargo signals.

Terrestrial AIS provides high-frequency updates but is limited to coastal zones. Satellite AIS (S-AIS) offers global coverage, including open oceans, but suffers from message collision and lower temporal resolution due to the wide footprint of satellite antennas. Modern data providers fuse both sources, using terrestrial data for high-resolution port activity and S-AIS for tracking deep-sea voyage progress. Understanding this fusion process is critical for assessing signal latency.

Non-traditional datasets sourced from outside standard financial filings and market data. AIS vessel tracking is a prime example, alongside satellite imagery, credit card transactions, and social media sentiment. These datasets generate unique trading signals uncorrelated with traditional price data.

• Used to predict commodity supply before official reports
• Requires specialized ETL pipelines for cleaning and normalization
• Often combined with traditional data via feature stores

A historical data snapshot preserving the exact state of a dataset as it was known on a specific past date. For AIS data, this means capturing vessel positions as they were broadcast, not corrected versions. Critical for eliminating look-ahead bias in backtesting.

• Prevents using revised coordinates in historical simulations
• Essential for accurate signal decay measurement
• Requires timestamped versioning at the record level

The computational process of identifying and merging disparate records that refer to the same real-world entity. In maritime contexts, this means linking a vessel's MMSI number, IMO number, and call sign across multiple AIS broadcasts to create a unified voyage record.

• Resolves duplicate or conflicting vessel identities
• Maps beneficial ownership to shell company structures
• Enables accurate port call detection and dwell time calculation

The prediction of the present or very near future state of an economic indicator using high-frequency, real-time data sources. AIS data enables nowcasting of global crude oil shipments, iron ore deliveries, and container throughput weeks before official customs data is released.

• Tracks real-time commodity flows to major ports
• Infers OPEC compliance from tanker movements
• Provides early warning of supply chain disruptions

The end-to-end tracking of data's origin, transformations, and movement through pipelines. For AIS data, lineage captures the journey from terrestrial receiver or satellite collector through decoding, deduplication, and geofencing to final analytical tables.

• Provides auditable maps for regulatory compliance
• Enables rapid debugging of signal anomalies
• Documents schema evolution across pipeline versions

A method of tracking and analyzing streams of information to derive conclusions in real time. Applied to AIS data, CEP engines detect patterns like dark periods (AIS transmission gaps), loitering events, and ship-to-ship transfers that may indicate sanctions evasion or illicit activity.

• Processes millions of position reports per second
• Triggers alerts on geofence breaches
• Correlates vessel behavior with weather data and port schedules