Predicting when rented equipment will be returned
A logistics optimization model for a leading provider of equipment for the construction sector
As one of the largest rental equipment providers in the Nordics and suppliers to thousands of construction sites, this client needed a system capable of predicting when lent out equipment was expected to be returned.
The goal was to optimize their vast logistics chain by minimizing unnecessary transport between their local shops and main warehouses.
In close collaboration with the company’s data scientists, we processed and analyzed their large database of customers, construction projects, and past rentals. As input, the model takes detailed characteristics of the equipment, the customers, geography, and season into account to estimate when equipment will be returned. We developed and assessed several regression and classification approaches.
The system was delivered as a Python service to be run on their internal infrastructure and integrated with the ERP system.