Making the Switch: Predictive Maintenance on Railway Switches

Download Slides

For a big Dutch railway maintenance company, Anchormen developed a predictive model on top of Spark that takes sensor data from all their switches and predicts whether a switch will break within the next 7 days, so our client can schedule a maintenance repair. Using the sensor data, we developed a decision tree based model for 7000 switches in the Netherlands, potentially impacting the lives of millions of travelers, with potential roll-out for the rest of Europe.

About Jeroen Vlek

Jeroen is a Spark and Scala enthusiast with degrees in Computer Science and Artificial Intelligence. After having done his thesis at the Fraunhofer Institute in Berlin, he worked in online gaming and consulting in Germany. In 2015 he and his wife moved to Groningen and Jeroen started as Lead Data Engineer at Anchormen. In his free time he enjoys techno music, computer games and anime.

About Chris Pool

Chris is a Data Scientist with a degree in Information Sciences and experience with machine learning and programming. Chris is Lead Data Science for Anchormen and is working on several big data projects. In his free time he likes to go cycling and play footbal.