My PhD thesis in a nutshell: data and models for multimodal mobility
What my PhD at the University of Bologna was about: combining Wi-Fi, mobile-phone and vehicle-sensor data with agent-based and optimal-velocity models, from the pedestrians of heritage cities to the traffic of Emilia-Romagna.
In May 2025 I defended my PhD thesis in Physics at the University of Bologna, Data analytics and predictive models for sustainable multimodal mobility in future smart cities. It was supervised by Armando Bazzani, with Mirko Degli Esposti as co-supervisor. This is a short overview for anyone who doesn’t want to read the full manuscript.
The question
Cities collect more and more data about how people move: Wi-Fi access points detect phones, mobile networks log connections, and roads are fitted with sensors that count vehicles. How can we turn these heterogeneous, noisy and incomplete sources into models that actually predict mobility, and that can support planning and real-time decisions?
Two scales, two case studies
The thesis follows the same approach at two very different scales.
Pedestrians in heritage cities. Historic centres such as Ferrara, Šibenik and Dubrovnik attract large flows of visitors in small, fragile spaces. Wi-Fi and mobile-phone data show where people actually go, and agent-based models reproduce and forecast those flows.
Vehicles at regional scale. For the Emilia-Romagna region, data from road sensors feed traffic simulations based on optimal-velocity models. These models capture how speed adapts to the vehicles ahead, and therefore how jams form and dissolve.
The common thread
Across both case studies the idea is the same: data on their own describe, models explain and predict. Combining the two gives tools that can help design infrastructure, manage crowds and traffic as they happen, and move toward more sustainable urban mobility.
📄 Full thesis (open access): doi:10.48676/unibo/amsdottorato/12316