2026-04-13T15:14:55Zhttps://recercat.cat/oai/requestoai:recercat.cat:10230/492182025-12-13T20:52:12Zcom_2072_6col_2072_452954
RECERCAT
author
Ferrer Sánchez, Josep
2021-12-15T11:50:55Z2021-12-15T11:50:55Z2021-07
Treball fi de màster de: Master in Intelligent Interactive SystemsTutor: Vicenç GómezDesigning and generating wiser policies for urban systems and infrastructures is a
challenge of paramount importance. Today, the cities that present the most successful
transport strategies are prioritising the movement of people, giving residents
and visitors a wider variety of attractive transport options while creating effective
ways to switch from private to public transport means. Understanding the use and
the impact of public infrastructures that facilitate mobility is crucial.
We consider a dataset of one year of activity in the form of car park occupancy in
the province of Barcelona. The data comprises ten different park station facilities
located close to train stations. We propose and analyze different models based on
statistical and mathematical techniques.
First, we analyze the occupancy recordings in different park station locations and
show that the activity is strongly coupled with the circadian rhythm, following a
24-hours cyclic pattern. Second, we perform a statistical characterization based
on different activity prototypes for days associated with similar activity. We show
that such a simple statistical model is enough to characterize accurately the car
park occupancy. Third, we propose a mathematical model that explicitly captures
the rise and decay of occupancy using a mixture of two different Gamma kernels.
We optimize the parameters for each park station independently and show that the
resulting model globally captures better the activity than the previous statistical
characterization, in terms of accuracy, simplicity (has less parameters) and interpretability.
Our results show that, despite the apparent complexity associated to public mobility
and use of car parks, very simple models motivated in intuitive principles are
sufficient to understand this dynamics to a large extent. Overall, our results can
facilitate the design of public policies to facilitate the mobility within Barcelona and
its surroundings, by providing a better understanding of how the citizens switch
between private cars and public trains.
© Tots els drets reservats info:eu-repo/semantics/openAccess
Car park
Car park occupancy
Prediction model
Gamma kernel
Modelling the use of car parks in the province of Barcelona
info:eu-repo/semantics/masterThesis