The demand for public transport: analysis of mobility patterns and bus stops

The availability of smart card data from public transport allows analysing current and predicting future public transport usage. As an essential part of public transportation, forecasting bus passenger demand plays an important role in resource allocation, network planning, and frequency setting. In this thesis, demand and offer of a public transport consortium located in north-western Italy have been taken into account. Also, demographical data about the customers and the land, and historical weather conditions of the area during the period of interest have been collected. Two main kinds of analyses of the demand have been performed: at first with respect to many possible features, such as type of day, hour of day, stop point, trip, kind of user (particularly in terms of age), kind of travel document. This has shown that three main kinds of days can be identified: working, half-holidays (like Saturday) and holidays; moreover, the demand shows a great variability across different trips, stop points and users and there are two daily time-slots with a peak of validations. Among the customers, the majority is represented by students. Then, the focus has been moved to the analysis of the stop points, which have been grouped according to the incoming demand, through a clustering algorithm. The proposed methodology has been developed in Python with the support of QGis software for geographical visualizations. The results may be the starting point for an efficient forecast of the demand, at any stop point and time interval. The implications of this thesis could be appealing for public transport operators, since forecasting the passenger demand is necessary to properly plan on-demand mobility services, increasingly being promoted as an influential strategy to address urban transport challenges in large and fast growing cities.