Master’s Thesis – Technische Universität Berlin
The logistics world still relies heavily on Bauchgefühl — gut feeling — when it comes to operational decisions. During my master’s studies, I was driven by the idea of bringing data-driven decision-making into this field, replacing instinct with insight.
In my thesis, “Improving Delivery Time with Predictive Analytics: A Case Study in E-Commerce,” I explored how machine learning could be used to predict shipment delays in cross-border e-commerce logistics. Using real shipment data from a European delivery platform, I developed classification models that identified which parcels were likely to be delayed — turning raw operational data into actionable predictions.
The project combined data engineering, machine learning, and analytics architecture design, resulting in both predictive models and a proposed alerting system for real-time logistics monitoring. This research showed that predictive analytics can transform reactive logistics processes into proactive ones — helping companies act before problems occur, not after.
View full thesis (PDF)