When I walked in for a cup of coffee at Xomnia to meet with Pieter over a year ago, I immediately felt at home. The somewhat unwieldy office, a great atmosphere, and open-minded people really made me feel welcome. After three cups of coffee, my job colored Xomnian-blue without even noticing that I’ve had applied somewhere.
After several very interesting interviews with potential clients, I started my adventure at ProRail. At ProRail I got the assignment to create a predictor for trespassers (spoorlopers in Dutch). A trespasser is defined as a pedestrian who walks too close to the tracks. When that happens, nearby trains need to slow down or stop, causing delays. Anticipating trespasser behaviour allows for measures minimizing the number of train delays, such as installing fences or repairing broken ones. For this prediction model, we used a combination clustering and classification techniques and a lot of open data. The predictor has become a dynamic application that displays daily hotspots through a dashboard. The Xomnia team and the junior development program supported me in both the technical and the process aspects of this project. Working according to the crisp-dm and the scrum method, involving the right people and technical additions of critical colleagues, was a big support. I’ve always experienced my Xomnia colleagues as inspiration to keep motivated when the going was tough, and to remain enthusiastic when successes were few. At the end of the week, there is always that beer to look forward to.
A few months after I started ProRail has begun setting up a DataLab. A DataLab is an environment where data scientists (like me) come together and have access to a great variety of tools and an abundance of computing power to tackle projects in a structured and efficient way. In addition to working on the implementation of this DataLab platform, I started the first official DataLab project; predicting railway switch failures. Meanwhile, the DataLab has officially opened and I am directly employed by ProRail. I hope to be one of the people who make the DataLab a success story.