Prediction system for the management of bicycle sharing systems

Abstract

Bicycle sharing systems are very common in urban areas; their goal is to improve citizens’ mobility around the city. The efficient management of bicycle stations is the greatest challenge for such systems and makes it difficult to satisfy the users’ demand for bicycles. To overcome this challenge, it is important to predict their use and to constantly monitor the number of bicycles available at stations, ensuring that they are distributed according to the demand in those places. In this work, we analyse the demand for bicycles per user and predict the routes users may travel to determine the possibility of predicting the behaviour of users. To this end, meteorological information and historical data on the use of the stations were incorporated into the system.