Understanding cycling mobility: Bologna case study

Abstract

Abstract Understanding human mobility in touristic and historical cities is of the utmost importance for managing traffic and deploying new resources and services. In recent years, the need to enhance mobility has been exacerbated due to rapid urbanisation and climate changes. The main objective of this work is to study cycling mobility within the city of Bologna, Italy. We used six months dataset that consists of 320,118 self-reported bike trips. First, we performed several descriptive analysis to understand the temporal and spatial patterns of bike users for understanding popular roads and most favourite points within the city. The findings show how bike users present regular daily and weekly temporal patterns and the characteristics of their trips (i.e. distance, time and speed) follow well-known distribution laws. We also identified several points of interest in the city that are particularly attractive for cycling. Moreover, using several other public datasets, we found that bike usage is more correlated to temperature and precipitation and has no correlation to wind speed and pollution. We also exploited machine learning approaches for predicting short-term trips in the near future (that is for the following 10, 30, and 60 minutes), which could help local governmental agencies with urban planning. The best model achieved an R square of 0.91 for the 30-minute time interval, and a Mean Absolute Error of 2.52 and a Root Mean Squared Error of 3.88 for the 10-minute time interval.

Keywords

• Cycling mobility
• Geo-referenced data analysis
• Big data
• Short-term mobility forecasting
• Machine learning