Characterising Annual Behaviour of Carsharing Users in Montreal

Carsharing is one of the emerging modes of transportation in the recent years, as a pointer to the fact that owning a private car is not all we need for the sake of travelling. Using k-means clustering and Principal Component Analysis, the purpose of this paper is to study the behaviour of the Communauto regular-service carsharing users in Montreal over a year (2014) and find the usage patterns in each cluster. Also, by having the Communauto customer features available, the characteristics of the customers in each cluster will be defined, which eventually ends in customer profiles. The k-means clustering results show that the Communauto regular-service carsharing users are divided into nine different clusters. Some of the clusters patterns are similar during the weeks and some are similar during a year, but none of them are similar during week and year. So, each cluster has a unique usage pattern over the weeks and the year. Furthermore, the resulting clusters are ordered from highly intensive users to the occasional ones based on the frequency of carsharing usage over one year.

Mots clés

Carsharing, User Behaviour, Clustering, Principal Component Analysis (PCA)