Skip to content

Data engineering consultancy in Rotterdam

We bring together sensor and ERP data from greenhouses, warehouses and transport. We’re building a data platform on Databricks for tomato producer Agro Care, based in Wateringen.

At Agro Care, all data comes together in a data platform with a data warehouse, ready for BI and data science

Agro Care, one of Europe’s leading tomato producers, is taking significant steps towards working in a more data-driven way. For the next step, a scalable and stable data environment, we’re building a data platform on Databricks that stores all data in one place and makes it accessible.

Read the case study

How do you put the climate control computer, harvest records and ERP on a single timeline?

The climate control computer takes continuous measurements for each section, harvest data is kept in a recording system, and orders and shipments are in the ERP, each with its own codes. We build pipelines that bring them together in one data platform, using the same identifiers for location, batch and customer.

Schedule a conversation

Measurement series, batches and orders in one platform used by both planners and controllers

Discuss your data sources

Sensor readings stored in raw form first

Climate, water and energy measurements are stored unchanged in a source layer. For Fokker Services Group, we modelled data using the medallion architecture.

ERP data delivered automatically

Orders, products and customer data come automatically from the ERP system. At Vivara, the data platform retrieves data daily from SAP and Magento, among other sources.

Master data with a history

If a variety moves to another department or a customer switches depots, the previous version is retained. That keeps last season’s figures accurate.

Ready for reporting and models

The platform provides tables for BI and analysis. That is the aim at Agro Care: to store all data centrally and make it easily accessible.

First, we get the keys for one source right, then connect the rest

  1. Codes and owners

    Which codes do the climate control computer, harvest records and ERP use for the same location, batch or customer, and who manages each source?

  2. One source through every layer

    We load one source in raw form, link it using the shared keys and deliver a table people can use to make decisions.

  3. Connect and manage

    Additional sources and sites reuse the same pipelines. Your own team then manages the platform, or we do.

Are our exports still enough, or is it time for a data platform?

Talk it through with us

A good fit if

You have sensor data from greenhouses, cold rooms or machines, alongside an ERP system and a planning system, and you want to compare multiple sites in the same way.

Not a good fit if

You need a data engineer to work temporarily alongside your own team: a secondment would be a better fit. If all your data comes from one system with good reporting, you do not need a platform.

What determines the scope

The number of sources and sites, how often sensors take measurements, how much historical data you include, whether your ERP system supports an integration or only exports, and who will manage the platform after go-live.

Who owns the data from your greenhouse equipment?

Suppliers of greenhouse and machine systems sometimes store data in their own cloud. The Data Act gives users of connected products the right to access that data. Your legal adviser can assess what this means for your contracts.

Supplier portal, data warehouse or lakehouse: where does your sensor data go?

Supplier portal

Choose this if you only need to view data by greenhouse or machine and don’t need to combine measurements with orders or costs. What the supplier provides is enough.

Data warehouse

Choose this if you’re mainly bringing ERP, planning and financial data together in fixed tables. Read more about building a data warehouse.

Lakehouse

Choose this if you need to bring long sensor time series together with ERP data and run models on them. Our advice: start with the tables people already use to make decisions.

As a Databricks and Microsoft partner, we build data platforms, including for Agro Care

Data Science Lab has worked with data and AI since 2016, from Westerdoksdijk in Amsterdam. As a Databricks Partner with certified consultants and a Microsoft Partner, we build data platforms for clients including Agro Care, Vivara and Fokker Services Group.

Book an introductory meeting

A list of your source systems and locations is enough for an initial conversation

Stef Stoelwinder

Stef Stoelwinder

Business Development

Stef is your first point of contact. In our initial conversation, we’ll discuss your sources, your locations and the question you still rely on exports to answer when making decisions.

Which systems and locations would you like to bring together on a single data platform?

Tell us about your source systems, sensors and sites. We’ll reply by email or arrange an initial conversation.

Enter a valid email address.

Let us know how we can help you.

Frequently asked questions

It builds and manages the pipelines and data platform that bring measurement data, ERP data and planning data together in one reliable place for reporting and models. A data engineer makes sure every source arrives on time, codes match up and errors are detected.

Greenport West-Holland, a network of businesses, public authorities and knowledge institutions in horticulture, runs programmes for robotics, digitalisation and fresh-produce logistics. Read more about the service on our data engineering page.

Agro Care is one of Europe’s leading tomato producers. According to its website, Agro Care Growers’ head office in Wateringen manages its growing sites centrally, including sites in the Westland, Middenmeer and Rilland in the Netherlands.

For Agro Care, we are building a data platform using a data warehouse and Databricks. The aim is to store all data centrally and make it easily accessible for BI and data science.

By using identifiers that mean the same thing in every source. The climate-control computer records a growing section, the harvest records a batch, and the ERP records a product and a customer. We document which code corresponds to which location, batch or customer in a mapping table, along with who maintains that table.

Only then can you compare a series of measurements with a delivery. Without that shared definition, the figures can appear correct until someone compares two sites.

Then you work with snapshots: each export is compared with the previous one, and only the changes move on. On Databricks, AUTO CDC FROM SNAPSHOT handles this; AUTO CDC processes a stream of changes if one is available.

Both can maintain a table as SCD type 1, which keeps only the latest version, or SCD type 2, which adds a new row for each version. Type 2 shows which customer or department a batch belonged to at the time.

Portbase is the Port Community System for Dutch ports and a subsidiary of the port authorities of Amsterdam and Rotterdam. The platform brings together, among other things, container notifications for inland transport and the status of incoming cargo.

You agree with Portbase which data you may access and how. We then treat it as another source alongside the ERP. Read more about data across the supply chain on our mobility and logistics page.

Often, yes, and since 12 September 2025 you have had a right to do so. According to the European Commission, the Data Act gives users of smart, connected products the right to access the data generated by their use and to share it with third parties.

Your legal adviser can assess what this means for your contract with the supplier. We set up data ingestion so it remains clear which data series comes from which installation.

Yes. In Rotterdam, we used cameras to capture the feeding process at Floating Farm, and we worked on the Botlekbrug for Rijkswaterstaat. These were AI projects; you can find the case studies on our data science firm in Rotterdam and AI consultant in Rotterdam pages.

Our office is on Westerdoksdijk in Amsterdam. After launch, we manage the platform through management and continuous improvement, or your own team takes over.

When your reporting comes from a single system and measurement data is only needed on the grower’s or operator’s screen. In that case, a report from the source system or the supplier’s portal is enough. A pipeline will not fix a sensor that measures incorrectly, either.

If you need someone to work temporarily within your own team, see hire a data engineer. If you want to compare providers, see our overview of data engineering firms.