Working in Data & Analytics: an opportunity to add real value for the users of ProRail

David has been working in KPMG’s Data & Analytics team since 2014. “What’s great about working here is the variety of projects and clients, which enables you to grow fast in your field.” In the Data & Analytics team you work on a wide range of assignments, sometimes for small, and sometimes for bigger clients. Recently for instance, David worked on a case for ProRail. And he’s keen to tell us about it.

The ProRail challenge

ProRail services over 7,000 kilometers of railway infrastructure in the Netherlands. This also means planning over 1,000 activities per year, and coordinating these in cooperation with various parties. Planning at this level is of course not something to be taken lightly. You need to consider the inconvenience for passengers and goods, the fluctuating costs of subcontractors, the cost of detours, and a multitude of other arrangements with stakeholders like transporters and municipalities. Large events must also be taken into account, to ensure that they don't suffer any inconvenience due to less or no train traffic. It is easy to see that manual planning is labor intensive, and that finding the optimal schedule is simply impossible without smart IT solutions.

To tackle these challenges, ProRail had a bespoke algorithm developed and implemented to optimize planning.

What did David and his team do?

We developed the TVP planner for ProRail. TVP stands for: ‘treinvrije periode’ (train-free period). The TVP planner helps to work out an optimal schedule under conditions that the user can select from, including for example a schedule that causes the least inconvenience for passengers. To do this, we worked together on complex topics.

Identifying and extracting appropriate data from the primary source systems was one of the first tasks in the project. We analyzed a lot of ProRail source systems, as well as external data. A thorough analysis of the available data also impacted the ultimate functionality of the application, since we had to work out what could be done with the available data. The team then wrote a data-transformation logic to link up data sources that had never been connected with each other in this way before. Together with a specialized consulting firm, the planning algorithm was implemented and a user-friendly interface was developed for the application.

The TVP planner application was implemented entirely on ProRail’s Azure Cloud platform. For this, we used container technology in combination with modern tools on the Azure platform. By linking it up to the ProRail network, the TVP planner application can be used like any other internal application, which is great for end users.

The various components of the application were developed in an iterative, agile process with short communication lines which ensured collaboration between all the parties involved.