Van Hout

I am proud of the platform that offers Van Hout customers insights

Client

Van Hout

The question

A stable IoT platform that facilitates a worry-free customer journey.

Approach

Making the existing platform maintainable and taking it into management.

Learnings

Reversed engineering from the front would have been better.

In the tail end of his master’s in Data Science and AI, Délano joins Blenddata as a Junior Data Engineer. After a six-week onboarding, Van Hout consultants and installers becomes his first major project.

“With OVVIA, Van Hout provides, among other things, sustainability reports on the indoor climate and energy consumption of buildings. This requires huge amounts of IoT data, sometimes every 15 minutes. The client wanted a stable data platform that they didn’t have to worry about. A platform had already been set up, but it was not ready for Van Hout’s future ambitions. We developed this platform functionally and we now manage it.”

From long to readable code

“The existing platform had a lot of Azure and Python code that was difficult to manage, especially as the number of data streams grew exponentially. In the first few weeks, I started analysing what all that code was doing. My next step was to make sure the code became more efficient and readable.”

Data engineering is the basis for data science and machine learning. So you better learn it well.

Délano Gaasbeek

Medior data engineer

You want code to read like a good book.

Délano Gaasbeek

Medior data engineer

Modern Data Stack

The toolbox for building a data highway

“At Blenddata, we use a standard toolbox based on open source techniques: Modern Data Stack. This includes tooling to access, transform and visualise data from sources. There is also tooling to set up the infrastructure that takes data from A to B. You try to configure that in code so that it is reproducible. Then you use a tool that controls orchestration, i.e. when and how often your pipelines run. If there is an outage at night, you can easily find that.”

A team of two gets the job done

Learning in practice how to programme properly, efficiently and legibly.

“As per our standard way of working, two consultants were put on this project. You help each other grow and take each other and the client on a journey. As this was my first project, I was paired with someone with more experience. In terms of technology and consultancy, I was able to link up with my colleague Rens and help build a stable platform. I learned a lot technically, how to write sustainable code, for instance. At university, you are not taught how to program properly, efficiently and legibly. This is covered in the six-week onboarding at Blenddata, but in practice it is even better and more fun, you see the added value immediately.”

Delivering value for the customer

“You first do a few days of preliminary research and then make a plan of action. Then comes the implementation phase. I am proud of the tangible result: a platform that delivers insights to Van Hout’s sustainability portal for customers. You immediately see the value it delivers. What I would do differently? In retrospect, it would have been better to look at what the end user sees from the front end, and engineer reversed from there.”

Sharing knowledge

Contributing to customer data literacy

“We ran a dashboarding and visualisation workshop at Van Hout. The employees got excited because they saw what had been invested in. They are now working with the platform themselves. Through this route, we contributed to data literacy at the customer.”

Someday I want to lead a large-scale greenfield project

“Very much I would like to lead, coordinate or design a large-scale greenfield project. Coming up with a working method and architecture from almost nothing. For now, it’s implementation. Inventing and designing is a step further, for that you need more knowledge. That’s the next step for me.

“What’s nice is that there is no minute-by-minute scrutiny of whether you make your hours. As long as you deliver the value promised to the customer, it doesn’t matter if you go 15 minutes early one time. The work culture at Blenddata is relaxed though and therefore suits me well.”