Taking Lausanne’s planned new tram line as a case study, Anne-Valérie Preto focused her master’s project in civil engineering on improving a computer model that’s used to forecast how people will use public transportation.
In 2018, Anne-Valérie Preto was a high-school student in Valais, in the physics and applied mathematics track, when she decided to attend EPFL’s open house. “I didn’t really know what I wanted to do after I graduated from high school,” she says. “But after visiting the civil engineering booths, I’d made my decision!”
As a bachelor’s student, she developed an interest in programming and optimization applied specifically to transportation systems. This entails creating algorithms that describe users’ behavior in order to better understand their preferences. For instance, transportation engineers look at what factors might prompt someone to drive to work or a leisure activity instead of taking public transportation.
It’s a highly complex field, which is something Preto enjoys. “Modeling people’s transportation choices provides important insight into how city residents get around, which is especially valuable today for designing environmentally friendly policies,” she says. After completing an exchange year in Lisbon, she began a master’s program in civil engineering with a specialization in transportation and mobility and a minor in computational science and engineering. This minor gave her key skills in data science, machine learning and statistics.
Analyzing preferences
For her master’s project, Preto decided to work with the SIMBA MOBi forecasting model developed by the Swiss Federal Railway (SBB). The model is used to plan urban routes and make transportation-related decisions based on qualitative modeling. Preto set out to improve the model by factoring in a variable that’s extremely hard to capture with algorithms: why some people prefer rail and light-rail systems to buses and other alternatives.
Preto used Lausanne’s new tram line, scheduled to begin operating in 2026, as a case study. She examined the “differentiation of modal preferences in public transportation,” or how people select between different kinds of transportation modes and map out their journey from point A to point B.
Her findings shed light on Lausanne residents’ preferences for public transportation. Some transportation modes compete with each other along a given trajectory: for example, Bus 1 (which runs between the Lausanne train station and EPFL) and the m1 metro line. Yet Preto found that different systems are often complementary. And the fact that residents generally prefer rail and light rail augurs well for the new tram line, which in turn is good news for the city.
Supervised by industry experts
SBB was involved directly in her research and implemented the enhancements to SIMBA MOBi that Preto suggested. This concrete application of her research findings was a real plus for Preto. “I was lucky to have great supervisors at SBB: Antonin Danalet and Joschka Bischoff, who work on the SIMBA MOBi team,” she says. “They helped me structure my research and keep moving forward throughout the semester.” Preto’s project received one of the highest grades in her class.
She now plans to do a PhD in civil engineering so that she can continue to help build better public-transportation models.
A typical day
Preto began her research by analyzing Swiss residents’ travel habits based on data from the most recent national transportation microcensus, where participants were asked to describe their typical day and what kinds of transportation they use. She combined these data with the results of the latest transportation survey carried out at the University of Lausanne and EPFL – two communities that will be affected by the new tram line. “We based our research on a ‘synthetic population’ of Swiss residents for whom we have general information on age, income and education level,” says Preto. “Then we ran decision models to determine which of these residents would own a car, a bicycle or a regional transport pass, for instance. After that, we used SIMBA MOBi to establish ‘typical days’ for these individuals based on their activities and where they live, including which modes of transportation they would use. For example, an EPFL student’s schedule could be a metro ride from Lausanne to the EPFL campus, then a tram ride after class to do sports or some other leisure activity, a bus ride to do some shopping and then back home.”
Light rail a popular option
The next step for Preto was a literature review to find out the difference among public transportation preferences. She discovered that rail or light rail is almost always an individual’s first choice, for several reasons: the higher frequency and reliability of these services, the greater comfort they provide – especially relative to buses – and the fact that their routes are easier to figure out. Surprisingly, she also found that a new light-rail line in a city can attract 20% more daily passengers than a new bus line, even at constant frequency and capacity.
Improving the model
Preto’s project included a discussion of how these preferences can be incorporated into SIMBA MOBi – which is no mean feat. “These factors are hard to integrate because the reasons why an individual might prefer taking the tram over the bus, assuming the same service level, are highly subjective and can’t be observed directly. For instance, someone might find a tram journey more pleasant than a bus ride,” she says. “Unfortunately, surveys of people’s stated preferences haven’t been conclusive. We have to rely on passenger counts, or statistics of how many passengers get on and off at each stop, and then calibrate the models to account for user preferences. Our goal is to be able to better represent the current state in public transportation systems and isolate the ‘light rail bonus’ more broadly, in order to explain modal preferences in Zurich just as well as in Lausanne.” As a last step, Preto tested the enhancements made to SIMBA MOBi by comparing the model’s output with existing public transportation data in Lausanne and Zurich.
Date : 2025-11-20
News source : EPFL.CH
Auteur : Sandrine Perroud
Anne-Valérie Preto, “Differentiation of Modal Preferences in Public Transportation: Analysis and Enhancement of the MOBi Model for Scenario Simulation in Lausanne,” master’s project supervised by Antonin Danalet (SBB), Joschka Bischoff (SBB), Negar Rezvany (EPFL Transp-or), Fabian Torres (EPFL Transp-or) and Michel Bierlaire (EPFL Transp-or), 10 January 2025.