Skip to content

A consistent and robust MLOps lifecycle for insights into travel behaviour

Read more

The client,

GVB’s core values are hospitality, connection and reliability. They guide its employees as they provide public transport in Amsterdam and the surrounding area. GVB makes Amsterdam accessible to everyone. With around 5,000 employees, agency workers and external contractors, it is one of Amsterdam’s largest employers – and ranks among the city’s five best employers.

Client GVB
Industry Mobility & logistics
Services AI Solutions
Partner The Cirqle
Tools & techniques Databricks, Azure DevOps

Our solution,

A segmentation model based on travel behaviour gives GVB more insight into how passengers move around. This helps GVB offer suitable travel products, for example. The model needs to stay up to date as travel patterns change over the years. It also needs to be implemented robustly to meet GVB’s data architecture standards. 

We established a consistent and robust MLOps lifecycle so the segmentation model can be deployed and maintained effectively in production. We also set up an MLOps workflow in Databricks, with several reusable templates.

It was very interesting to improve the Databricks workflow for machine learning projects by establishing a robust MLOps lifecycle. This has laid a solid foundation for setting up future projects in a similar way.
Koen Koopen Senior Data Scientist - Data Science Lab

What do you want data to do?

Enter a valid email address.

Let us know how we can help you.