One step ahead of delays at KLM

To reduce the impact of delays on the operation at KLM, a Xomnia X-Force team designed and implemented one predictive model and revised and improved another, going from start to production in under three months.

To reduce the impact of delays on the operation at KLM, a Xomnia X-Force team designed and implemented one predictive model and revised and improved another, going from start to production in under three months.

The main challenge that we see with data science at companies nowadays is identifying the right problem that actually adds value to the business. The second challenge is developing a mature way of working with the right skill set to bring a concept all the way to production.
Robin van den Brink, X-force analytics translator


KLM Royal Dutch Airlines is a Dutch airline based out of Schiphol that flies to more than 160 destinations worldwide. Established in 1919, KLM has recently been increasing their efforts to become more data-driven in order to remain a front runner in a competitive industry.

As an airline business, delays have a huge day to day impact on operational processes in the company. For example, ground processes (like cleaning and fuelling of the plane), baggage services and customer satisfaction could be severely influenced. But what if you could predict these delays in advance so you could try to mitigate the loss, and have a team to help you solve it?

An X-Force team – formed by an analytics translator, two data scientists and two data engineers –  was assigned to work on researching the repercussions of delay at KLM Royal Dutch Airlines. They looked at different use cases to determine which solution would have the most effective action for day-to-day operations at KLM.


The X-Force team found two viable areas to focus their attention on. The first was improving a preexisting delay prediction model, which predicts delays the day before a flight. The second was to create a new prediction model that supports an operational tool to streamline the operations at the day of flight.

Using information about every single flight, involving the conditions at Schiphol Airport and foreign airports including the weather forecast, aircraft load, and historic delays, a deep neural network has been trained to predict the delay of a flight.

The outcome of both developed prediction models are written to one of the data platforms at KLM. The predictions are consumed by several data-driven tools that are developed by KLM to support operations. During the development process, a dashboard was written in Python and Dash, and deployed to a Kubernetes cluster. This was to help end users to understand the predictions and verify the outcome based on years of domain expertise.

For optimal results, the data-driven tools not only need predictions but also confidence in all predictions. This means that the deep learning architecture is customised to generate a distribution of the expected delay. Using a probability distribution instead of a single prediction value, the optimisers can minimise the expected cost while at the same time reduce the impact on scheduled processes by avoiding unnecessary rematching of airplanes to planned flights.

Just like the dashboard, the model is also deployed to a Kubernetes cluster. Using the Atlassian CI/CD stack, improvements to any component of the model can be tested and deployed in a matter of minutes.


At the time of writing, two prediction models produce a combined 18,000 predictions per month to support the day to day operations of KLM and feed their operational optimisation tools.

Within 3 months, the X-Force delivered a revised, retrained and significantly improved the pre-existing model and deployed it in the new Kubernetes architecture. Plus, a new model was created in three sprints that is capable of continuously predicting delay for all the (upcoming) flights during that day.

KLM is now better able to mitigate the consequences of delay and act proactively which prevents the forced cancellations of flights. They can proactively swap planes, reschedule flights, rebook passengers, and prevent people from having to stay overnight in Schiphol.

Indirectly this means happier customers (though they might not be aware of it), and a future-proof system that can cope with delays without it having a huge impact on the day-to-day operations.

KLM has asked the X-Force team to continue working on the two predictive models and expand the scope on all of the processes around aircraft to identify opportunities to make the operation more predictable. The X-Force way of working has become the new standard way of working for the entire Prediction team of KLM. This was realised with a close intertwined collaboration between the X-Force and the Prediction team, with several workshops, presentations and hands-on support.