Modeling cyclist behavior using entropy and GPS data.

This paper analyzes the relation between Shannon's information entropy and trips made by cyclists in an urban context. Such entropy indicator depends on people distribution across a territory in a specific time. Its evaluation and temporal evolution have multiple applications in transport planning, from the specification of demand distribution models to the calibration and validation of transport demand models to be used where transport data are scarce. This feature is crucial in transport planning, as the lack of robust data to estimate travel demand is still a common challenge, especially for cycling mobility. The entropy indicator has been applied to the cyclists' case in the Italian city of Bologna, using approximately 165,000 GPS traces recorded from April to September 2017. The results are encouraging because the entropy indicator is able to explain the evolution of the agents' distribution throughout a day across the territory. However, it is suggested that future research may further deepen the concept to test its transferability, and to better establish its sensibility to other case studies. [ABSTRACT FROM AUTHOR]