Skip to content

Data engineering consultancy in Eindhoven

We bring machine, sensor, quality and ERP data together in pipelines and a lakehouse, as we’re doing in the scalable data environment we’re building for producer Agro Care.

A scalable, stable Databricks data environment where Agro Care stores all its data centrally

Tomato producer Agro Care is taking significant steps towards working in a more data-driven way. We are building a data platform with a data warehouse that makes its data available for BI and data science. Agro Care is not based in Brainport; this case study shows how we provide a producer with a platform.

Read the case study

How do we bring machine logs, measurement results and ERP orders together?

At high-tech and manufacturing companies in Brainport, machine data is on the shop floor, measurement results are in a quality system and orders are in the ERP. To trace a rejected product back to a machine or batch, you have to search three systems. Pipelines establish that connection each day.

Book a call

From machine to order in one platform, with keys connecting batches, production lines and customers

Discuss your data flows

Machine data without disrupting the production line

We read machine and sensor data from a historian, an export or a gateway on your own network, without interfering with the control system.

Measurement results linked to orders and batches

We link measurements and rejected items from quality control to the order number, batch and machine, so you can trace a deviation back to the production step.

From raw to usable data, layer by layer

Raw, cleaned and business-ready data sit in separate layers. For Fokker Services Group, we model data from many sources using the medallion architecture.

Releases and infrastructure as code

Pipelines and environments are version-controlled and go live through a standard release pipeline, as we did for SNG when we automated releases through the DTAP pipeline.

First, one production line from machine to reporting table, then the rest of the factory

  1. Sources and keys

    Identify the machines, systems and exports, and the key that links each order, batch and machine.

  2. One line end to end

    Take one line or product group from source to table, with checks for data freshness, completeness and schema.

  3. Expand and hand over

    Connect the next lines and sites, set up monitoring, and leave management to your own team or to us.

Is our factory data ready for its own platform, or should we start with something else?

Talk it through with us

A good fit if

You make products across multiple production lines or sites, your machines already log data, and quality, planning and finance teams want to see the same figures for each order, batch or machine.

Not a good fit if

Everything is in one ERP system, with reports that meet your needs. Or you mainly want an additional engineer to work alongside your team; staff augmentation is the better route for that.

What determines the scope

The number of production lines, machines and sites; how often sensors take readings; whether data sources offer a connection or only exports; how much historical data needs to be included; and who will manage the platform afterwards.

Can shop-floor data leave the factory?

Machine networks are often separate from office IT. Together with your IT and OT teams, we agree which data goes to the platform and by which route, or whether to aggregate it locally.

The supplier’s historian, a data warehouse or a lakehouse: where do you store your machine data?

Historian or MES

Choose this if you mainly look at trends for each machine and no one needs to compare that data with ERP, planning or quality data.

A data warehouse

Choose this if you’re dealing with orders, inventory and finances, and a summary of machine data per shift or order is enough.

A lakehouse

Choose this if raw measurement series, quality data and ERP data need to come together for reporting and models. We recommend this for multiple production lines. Read about a data lakehouse.

From Amsterdam, we build data platforms for manufacturing companies as a Databricks and Microsoft partner

Data Science Lab has been building data foundations since 2016. As a Databricks Partner and Microsoft Partner, we’ve developed data platforms for Vivara, Fokker Services Group and Agro Care, among others.

Book an introductory meeting

An overview of your machines, systems and exports is enough for the first conversation

Stef Stoelwinder

Stef Stoelwinder

Business Development

Stef is your first point of contact. In our first conversation, we’ll discuss which machines and systems provide data, where that data is currently stored and which line to start with.

Which machine, system and export need to be connected first in your organisation?

Tell us which sources you have on the shop floor and in your office IT systems. We’ll reply by email or suggest an initial conversation.

Enter a valid email address.

Let us know how we can help you.

Frequently asked questions

It builds the data flows that reports and models rely on: pipelines from machines, quality systems and ERP, a lakehouse or data warehouse, checks and access controls. In Brainport, this often involves high-tech and manufacturing companies with many production lines, product variants and suppliers.

If you’re looking for a data engineer or a data platform in Eindhoven, we provide a team to build the whole system and hand it over. Read more about our data engineering service.

We only read data; we never write back to the control system. This is often done through the historian or MES that already collects the data, or through a gateway on your own network that only makes outbound connections.

Your IT and OT teams decide which approach is permitted. We show you what each approach requires in terms of network access, permissions and management.

Using an identifier found in every system: an order number, batch number or serial number, supplemented by the machine and timestamp. If an identifier is missing from a source, we work with your production team to establish a consistent rule for deriving it.

Only when that link is reliable can a report show which production line or material is behind a deviation.

It organises data into layers: bronze for raw data as it arrives, silver for cleaned and linked data, and gold for tables ready to use in reports or models. This lets you trace errors back to their source.

At Fokker Services Group, we bring data together, clean it and model it using this architecture, with data quality checks built in.

We build on Databricks, Microsoft Fabric or your existing environment, choosing what fits your IT landscape. For Vivara, we used Azure, Databricks, dbt and Terraform, among other tools.

We define pipelines and infrastructure as code and provide documentation. That way, your own team or another provider can take over the platform.

We agree on that in advance. Your own team can take over, with documentation and a handover. Or we can continue to monitor and expand the platform with new production lines, machines or data sources.

Read about how this works in management and continuous improvement.

For models based on your machine and maintenance data: data science consultancy in Eindhoven. For reporting: Power BI specialist in Eindhoven. For AI that runs in production: AI consultancy in Eindhoven.

If you first want to find out where data could help your factory most, read about our AI consultant in Eindhoven. The same service for large organisations with many data sources is covered under data engineering consultancy in Utrecht.

If a single ERP or MES already answers your questions, or if machine data is only viewed machine by machine and never compared with orders or quality data. A platform also needs managing.

We’ll say so in the first conversation. Sometimes a better export or a consistent definition in the ERP is enough.