AI consultant in Rotterdam
We first test whether a model adds value using your own ship, sensor or planning data, as we did for Rijkswaterstaat at the Botlekbrug and Spijkenisserbrug.
We simulated traffic congestion caused by a bridge opening for Rijkswaterstaat
Together with the NGinfra partners, we investigated whether data could improve the operation of the Botlekbrug and Spijkenisserbrug for shipping, freight trains and road traffic; the first proof of concept, using AIS data, a congestion model and an LSTM, showed that it could.
Which decisions about bridges, quays or berths are still based on gut feeling?
A planner can see from AIS data that a seagoing vessel will arrive later than expected, but still adjusts crane schedules, driver assignments and train paths by hand. A model that combines vessel movements, schedules and traffic levels can calculate in advance how a decision will affect the rest of the chain.
Advice based on data from ships, sensors and camera footage
A proof of concept first
Your own historical data shows whether a model adds value. For Rijkswaterstaat, we used it to simulate the effect of opening a bridge on traffic congestion.
Figures to inform port policy
For Port of Amsterdam, we made historical berth use clear, so policymakers can use the figures to decide whether more or fewer berths are needed.
Cameras where nothing is measured
At the Floating Farm in Rotterdam, we installed cameras ourselves and used computer vision to capture the feeding process.
Someone to take ownership after the pilot
A model for bridge openings or activity along the quay becomes outdated when shipping patterns change. We agree in advance who will monitor, retrain and manage it.
The bottleneck between waterways, roads and rail determines where we start
-
Bottlenecks and partners across the chain
Which decision creates a bottleneck at a bridge, quay or in planning, and which parties are affected or provide the data?
-
Testing against historical data
A proof of concept using historical AIS, sensor or planning data shows whether a model adds value.
-
A proposal in the existing system
The model makes a proposal in the system the operator or planner already uses. They make the decision themselves.
Is your challenge in ports, water or food ready for AI?
A good fit if
You have historical data on vessel movements, sensors, schedules or images, and a recurring decision that currently relies on experience, such as opening a bridge, allocating a berth or deciding the order of operations at a terminal.
Not a good fit if
The data is held by a partner in the supply chain who does not share it, or the decision comes up so rarely that a fixed rule or a good dashboard is enough. If so, we’ll recommend that instead.
What the scope depends on
The number of parties in the port or supply chain providing data, whether data needs to arrive in real time, how much historical data is available, and whether the project remains a proof of concept or goes into production.
Who takes ownership across the chain?
The uncertainty is rarely about the model. More often, it is about who shares the data and who manages it when more organisations benefit from it. We agree on this before development begins.
Sensors, schedules or camera footage: which data should you put to work first?
Sensors
Choose this if measurements are already being collected over time, such as AIS positions, water levels or bridge monitoring. Our advice for port and infrastructure questions: that historical data usually already exists.
Schedules
Choose this if the opportunity lies in sequence and timing: which bridge, crane or driver is needed when. You’ll also need data on what actually happened; see optimising schedules with AI.
Camera footage
Choose this if what you want to know isn’t measured anywhere, such as feed left uneaten. Allow for labelling work and GDPR agreements; read more about computer vision.
Experience in waterways, ports and urban farming, based in Amsterdam
Data Science Lab has worked with data and AI since 2016 and is a Databricks and Microsoft partner. Through our Data for Good initiative, we worked for Floating Farm. We can meet at your place, at ours or online.

One decision on the water, on the road or at the quay is enough for an initial conversation
Stef Stoelwinder
Business DevelopmentYour first point of contact. In our first conversation, we’ll discuss one decision in your chain, the data involved and who manages it.
E-mail info@datasciencelab.nl
Phone +31 (0)20 3636 724
Which bottleneck between ship, quay and customer will you tackle first?
Describe the bottleneck and the data you already have. We’ll reply by email or arrange an initial conversation.
Message sent
Thanks for your message. We’ll be in touch as soon as possible.
Frequently asked questions
AIS (Automatic Identification System) is the system ships use to automatically broadcast their position, speed and voyage details. According to Rijkswaterstaat, it has been mandatory since 2016 for commercial vessels and recreational craft longer than 20 metres on waterways governed by the Inland Navigation Police Regulations. You can use this historical data to work out when ships reach a bridge, lock or quay.
AIS data was one of the sources for the Botlekbrug and Spijkenisserbrug. Read about how data helps planners spot disruptions in the chain earlier in data and AI for logistics.
With a congestion model, which calculates how a queue forms and clears when the road is temporarily closed. In the Botlek case study, we also used AIS data and an LSTM, a neural network that learns patterns in time series.
The proof of concept showed that you can simulate the effect of a bridge opening on traffic congestion. The long-term goal is a system that suggests an opening time with the least overall disruption to the bridge operator. The operator makes the decision.
Start by agreeing who provides which data, how often and under what terms. Decisions by operators across the region affect one another: the province of Zuid-Holland and Rijkswaterstaat agreed that the Hartelbrug and Spijkenisserbrug would never open at the same time.
In the Botlek case study, we worked for Rijkswaterstaat with the partners of NGinfra, a knowledge platform that brings together organisations including Havenbedrijf Rotterdam, ProRail and Rijkswaterstaat. A shared data and AI platform only makes sense once those agreements are in place.
Havenbedrijf Rotterdam collects real-time data from sensors in the port on infrastructure, water and air through an IoT platform. Rijkswaterstaat continuously monitors the operation of the Botlekbrug. These measurements provide the basis for predictive maintenance and for forecasting water levels or congestion.
A series of sensor readings only becomes meaningful when you can link it to faults, maintenance and events. Making those connections often takes more work than building the model itself.
Yes, if what you want to know is visible and is not measured elsewhere. At the Floating Farm in Rotterdam, the cows’ feed partly consisted of by-products from the city, such as spent grain from Rotterdam breweries. We installed cameras and built a computer vision model to make the feeding process visible and help prevent feed waste.
If visitors or staff can be identified in the footage, the GDPR applies. Decide in advance which footage you will keep and who can see it.
For the Botlekbrug, the goal is a system that suggests an opening time to the operator; the operator decides. Where a system acts autonomously, oversight remains in place: the Maeslantkering closes fully automatically, but when high water is forecast, a Rijkswaterstaat team monitors the process and intervenes if necessary.
Annex III, the AI Act’s high-risk list, includes AI used as a safety component in critical digital infrastructure, road traffic and the supply of water, gas, heating or electricity. Your legal or compliance team must assess whether your application for a bridge, flood barrier or road falls into this category. You can find out which obligations apply and when in the AI Act and its timeline.
No. Data Science Lab has one office, on Westerdoksdijk in Amsterdam. We can meet at your office, at ours or online.
For the Botlek case study, we worked with the NGinfra partners on two bridges in the region, and at the Floating Farm we installed the cameras ourselves.
What will you read first: the approach, logistics or computer vision?
- AI consultancyThe approach: how we identify and develop AI opportunities with organisations across the Netherlands.
- Data and AI for logisticsThe sector: how data helps planners and supply chain partners deal with disruptions in transit.
- Developing computer vision solutionsThe technology behind camera footage, from labelling to retraining after launch.