MaritimeTransOceanic Shipping (name changed)

Global Fleet Route Optimisation

Commercial cargo vessel approaching port

Case study. Client names are changed for confidentiality. Results are measured against each client's pre-project baseline.

18%

Bunker Cost Reduction

Saving over $12 M annually across the fleet.

94%

On-Time Arrival Rate

Up from 82% — a 12-point jump in schedule reliability.

45 k t

CO₂ Avoided Year 1

Equivalent to removing ~10,000 passenger cars from the road.

11 hrs

Avg. Steaming Time Saved

Per voyage, freeing capacity without adding a single vessel.

Case studyClient names are changed for confidentiality. Results are measured against each client's pre-project baseline.

The Challenge

TransOceanic Shipping had 150 vessels on static routing plans built from week-old weather charts. They could not react to real-time ocean-current shifts or last-minute port closures. The result was an average of 11 hours of wasted steaming per voyage and ETAs that nobody trusted.

Our Approach

1
Ingested AIS vessel-position data, NOAA GRIB2 weather grids, and port congestion APIs into a unified streaming pipeline.
2
Built a Digital Twin per vessel using actual hull-resistance and engine-load curves to simulate fuel consumption against any candidate route.
3
Trained an XGBoost ensemble on 8 years of historical voyage data to predict optimum speed and heading adjustments in 4-hour windows.
4
Deployed a Power BI-embedded captain's dashboard with visual route overlays, risk scores, selectable alternatives, and one-click acceptance.
5
Integrated automated ETA notifications to charterers via API, closing the loop from model output to commercial communication.

The Solution

We built a routing system that continuously re-evaluates each vessel's optimal path. It ingests live satellite weather, wave-height, and current data every 15 minutes, feeds them into per-vessel physics models, and surfaces ranked route recommendations to bridge officers via a custom Power BI dashboard. Captains keep final authority — but they have data-backed alternatives instead of gut feel.

Tech Stack

Data Engineering

Python (Pandas, NumPy)

Core data wrangling, route-simulation logic, and batch preprocessing of GRIB2 weather files.

Apache Kafka

Real-time streaming of AIS position events and weather-grid updates at sub-minute latency.

PostGIS (PostgreSQL)

Geospatial storage of historical voyage tracks, waypoints, and port-boundary polygons.

ML / AI

XGBoost

Gradient-boosted ensemble predicting optimal speed-heading combinations per 4-hour waypoint.

Integration

NOAA / Copernicus APIs

Hyper-local ocean current, wave-height, and atmospheric-pressure data ingested every 15 minutes.

Infrastructure

AWS SageMaker

Managed model training, versioning, and blue-green deployment with auto-scaling inference endpoints.

Docker / Kubernetes

Containerised microservices for the routing engine, ensuring zero-downtime updates at sea.

Visualisation

Power BI Embedded

Captain-facing route dashboard surfaced inside the ship's existing ECDIS interface.

Face a similar challenge?

Send us your challenge. We will reply to arrange a conversation and scope the next step.

Start with the decision

Turn one operational challenge into a measurable data plan.

Tell us about your challenge. We will reply to arrange an assessment and identify a practical route to delivery.

Request a data assessment