Skip to content

Data science consultancy in Den Bosch

We build models that calculate the impact of a decision in advance using your own data. For Rijkswaterstaat, we simulated how opening a bridge affects traffic congestion.

With NGinfra, we showed Rijkswaterstaat that the impact of a bridge opening on traffic congestion can be simulated

Together with the NGinfra partners, we explored whether data could improve the operation of the Botlekbrug and Spijkenisserbrug for shipping, freight trains and road traffic. The proof of concept showed that it could.

Read the case study

What happens if we schedule things differently, and can we calculate the outcome first?

A different time, a different order or an extra shift: you usually only see the effect afterwards. A model using your historical and current data calculates it in advance, so you can choose the option with the least impact.

Schedule a conversation

From historical data to a model that assesses decisions in advance

Discuss your challenge

Simulating impact

For Rijkswaterstaat, we used a congestion model and AIS data to simulate how opening a bridge affects traffic queues.

Signals from sensors

For ProRail, we used machine learning to detect trains, track components and wheel defects from fibre-optic measurements.

Usage in numbers

Together with Port of Amsterdam, we developed a tool that makes historical berth usage clear.

A recommendation, not a black box

The model shows which option it recommends and why, so the planner or operator can make the decision.

Recreate what happened before modelling new scenarios

  1. Data and decisions

    Which decision you want to improve, what data is available and what you can and cannot change.

  2. Recreating the past

    The model simulates a known period so you can see whether it accurately reflects reality.

  3. Modelling alternatives

    Different timings or sequences are modelled to give decision-makers the best possible support.

Can we model our planning using data?

Assess it in a conversation

A good fit if

You repeatedly make the same kind of decision about timing, sequence or capacity, have historical data on what happened afterwards, and want to see the impact in advance.

Not a good fit if

The decision rarely comes up, or there is no data on the effects of previous choices. In that case, measuring first is more useful than building a model.

What determines the scope

The number of sources and parties involved, how up-to-date the data needs to be, whether it is a pilot or a system in use, and who manages it.

Do we start with a pilot?

Often, yes. As with Rijkswaterstaat, a Proof of Concept first shows that it can work before a system is built for daily use.

A rule of thumb, a spreadsheet or a model based on your own data?

A rule of thumb

Choose this when the situation is stable and experienced people can judge the effect reliably.

A spreadsheet

Choose this when only a few factors are involved and you can track them yourself. We recommend this as long as the relationship remains simple.

A model based on your data

Choose this when many factors combine to determine the effect, such as timing, how busy it is and the order of events, and you want to compare different options.

We modelled options for rail, water and ports

Rijkswaterstaat, ProRail and Port of Amsterdam have already worked with us. Data Science Lab has worked with data and AI since 2016, as a Microsoft Partner and Databricks Partner.

Book an introductory meeting

We start by discussing which decision keeps coming up and what data is available

Stef Stoelwinder

Stef Stoelwinder

Business Development

Stef is your point of contact. In our first conversation, we discuss which decision keeps coming up and what data is already available.

Which planning decision would you like to run the numbers on first?

Describe the decision and the data you have. We’ll reply by email or arrange an initial meeting.

Enter a valid email address.

Let us know how we can help you.

Frequently asked questions

It builds models that predict outcomes, calculate scenarios or recognise patterns, and makes sure they are used in day-to-day work. This page focuses on organisations in 's-Hertogenbosch and Brabant that want to calculate the impact of planning decisions in advance.

You can find other types of models under AI applications.

Our office is in Amsterdam. We work mostly remotely with organisations in Den Bosch, and hold sessions with planners and operators on site.

The case studies on this page are from elsewhere; they show the type of model we build.

It is a model that reproduces how a situation develops, so you can calculate what happens if you change something. For Rijkswaterstaat, we used one to simulate the impact of a bridge opening on traffic congestion.

We used a congestion model, an LSTM model and AIS data from vessels.

Historical data about the situation and the decisions made, such as timings, traffic levels and waiting times, preferably over a longer period. The more variation there is in past decisions, the better the model can learn their effects.

If that data is missing, the first step is to start measuring.

By having it recalculate a known period and comparing the results with what actually happened. Only when it accurately reflects the past do you use it to assess new scenarios.

We repeat that check when the situation changes.

Yes, if current data is available, such as measurements or positions. The model can then suggest the best time, while the operator makes the decision.

You can read how we keep models running in production under MLOps.

That page is about advice: where AI fits in your organisation and what to tackle first. This page is about building a model that calculates the effects of different decisions.

See AI consultant in Den Bosch.

When a rule of thumb or a simple spreadsheet gives the same answer, or when nobody will use the results to plan differently.

If you mainly want to look ahead at demand or traffic levels, read about forecasting.