Phase: Requirements & System Design (SDLC Phase 1-2)
Core Stack: Python, Docker, FastAPI, Leaflet.js, Pandas/GeoPandas
The Intelligence Problem
Vessels evading Russian sanctions often form a "shadow fleet." Their primary tactic is the deliberate disabling of transponders to conduct illicit ship-to-ship transfers or approach embargoed ports unseen. Current monitoring is either manual, reactive, or lost in the noise of global shipping.
This project designs a targeted system to automate the detection of this specific deceptive behavior by focusing on high-risk vessels exhibiting suspicious gaps in high-risk zones.
System Architecture & Methodology
The platform is designed as a containerised, three-stage intelligence pipeline to transform terabytes of public AIS data into prioritised, investigation-ready alerts. The architecture follows a data flow from ingestion to red flag.
Stage 1: Static Risk Profiling
The system first filters global tanker traffic using known shadow fleet markers to create a targeted watchlist, reducing data noise by over 90%.
- Filters: Vessels >15 years old, flying high-risk flags (e.g., Cameroon, Comoros), linked to opaque shell companies.
- Outcome: A focused watchlist of 500-1000 high-probability vessels from a pool of ~15,000.
Stage 2: Behavioral Analysis & Gap Detection
Continuously monitors the watchlist for the key behavioral signature: deliberate AIS disappearances not explained by port visits.
- Core Logic: Flags AIS signal losses exceeding 12 hours occurring outside a 20km port radius and known coverage blind spots.
Stage 3: Contextual Scoring & Alert Generation
Each potential event is scored using weighted factors to create prioritised, actionable intelligence.
- Risk Factors: Gap duration (40%), geographic zone risk (30%), vessel's history of gaps (20%), and owner opacity (10%).
- Output: A structured JSON alert containing the vessel's last known position, a calculated risk score, and a link to its full forensic dossier.
Technical Design & Deliverables
The project is structured as a modular, documented system ready for implementation, reflecting production-grade planning.
Repo Structure
/backend/– Core detection engine (Filtering, Gap Detection, Alert Generation modules)./frontend/– FastAPI server & Leaflet.js dashboard for alert visualisation./config/– High-risk zone GeoJSON, port lists, vessel watchlists./docs/– Complete architecture, algorithm, data schema, and deployment specs.
Investigation-Centric Outputs
- active_alerts.json: Real-time feed of current high-risk gaps.
- Vessel Dossiers: Per-vessel JSON files with static profile, behavioral history, and past incidents.
- Interactive Dashboard: Map interface to triage alerts and access dossiers in one click.
Project Status & Development Roadmap
This project is in the initial phases of a structured Software Development Life Cycle (SDLC), focusing on comprehensive requirement gathering and architectural specification.
- Phase 1 (Current): Requirements Analysis & System Design – Core concept validation and architectural specification.
- Phase 2 (Planned): MVP Implementation – Development of the core detection engine and backend API.
- Phase 3 (Planned): Frontend Dashboard – Development of the interactive Leaflet.js visualisation interface.
- Future Roadmap: Includes live testing, calibration, and design for advanced pattern detection (spoofing, identity switching).
Links & Documentation
This project is developed transparently. The full technical specification and source code are available.