Vessel Prediction Model · Baltic & North Sea

Stay one step ahead of your customers

VPM predicts a vessel's next ports of call — and what it will need when it arrives.

Live AIS Baltic & North Sea Measured accuracy
The problem

Selling to ships is a timing business.

Today, the timing arrives by phone call, by agent, by luck — and often after a competitor has already quoted.

At random Late After the competitor
What it does

Turn "find out by chance" into "know in advance".

Know when a vessel reaches a port  →  know when a need arises  →  know when to make the call.

01 · Predict

Next ports of call

Where each vessel is heading next, plus the second and third calls ahead.

02 · Estimate

Fuel, cost & CO₂

How much fuel the leg takes, which grade, what it costs, and the emissions with it.

03 · Reach

Operator contacts

A contact for the operator, so the prediction is something you can act on.

How it works

From a live signal to a decision.

Four steps, running continuously.

Step 1

Live AIS in

Positions from open AIS feeds, continuously.

Step 2

Learn the pattern

How each vessel has moved between ports, and where it says it is going now.

Step 3

Predict the next ports

Weighing pattern against declaration, per vessel — plus two calls further out.

Step 4

Estimate the need

Fuel, cost, arrival and departure windows, along the real sea route.

Measured, not guessed. Every accuracy figure comes from a chronological back-test — the model only ever sees history from before the call it predicts. Read the full pipeline →
Accuracy

Not a guess. Measured.

77.9%
Next-ports Top-1 accuracy
events-weighted, tested fleet
76.2%
Same measure
strict chronological hold-out
Accuracy is higher when a vessel declares its destination and lower when it doesn't — VPM shows you which case you are looking at, rather than averaging the two. Accuracy & methodology →
Vessels converging on a port along predicted routes
The Port View

Ask it the other way round.

Pick a port. See the vessels heading toward it over the coming days — each with its load state, its likely need, its departure window, and a contact. Not a database of ships: a short list of who to call, and when.

How the Port View works →
Who it is for

One port is one market.

The same tool, sold once per port, to whoever is timing a need there.

One engine, more layers

Fuel was the first proof, not the last layer.

Fuel

When, where, how much, what grade.

Live
Every port, multiplied by every layer, is a market. Each one built to the finish before the next begins. See the roadmap →
Network of ports and shipping routes across the Baltic
Coverage & data

Baltic and North Sea, live.

The engine is geography-independent — it works anywhere there is AIS and shipping-lane data. Built on open AIS and public datasets, with a paid feed as backup. VPM sells decisions, not raw AIS.

~460 ports live History from 2020 ~9,000 Baltic vessels a year — the market
Coverage & data in detail →
Documentation

The working reference.

How it is built, how accurate it is, what data goes in — written straight, with the weak spots left in.

All documentation →
Questions

The ones people ask first.

Is this legal?

Yes. VPM uses open AIS and public datasets, with a paid feed as backup, and does not redistribute raw AIS or any provider's dataset — it sells derived decisions. Data & sources.

How accurate is it, really?

The first of the predicted next ports is right about 77.9% of the time across the tested fleet, on a strict chronological back-test — higher when a vessel declares its destination, lower when it doesn't. The full breakdown.

Which regions does it cover?

The Baltic and North Sea are live. The engine is not region-specific, but a port is only useful once enough of its traffic is tracked. Coverage.

What does it cost?

Pricing is not published — it is structured per port and per capability layer. Get in touch with your port and use case.

More questions →

See it for your port.

Tell us where you work and how you handle timing today.