La Ciencia de Bicing

The price of oil, traffic congestion, and environmental concerns have lead city governments and citizens alike to seek out alternative means of transportation. One such alternative transport called "shared biking," which provides low cost access to communal bicycles in a city, has experienced renewed interest in many cities.

Recent deployments such as those found in Barcelona, Paris, and Lyon have been enormously successful, each boasting over 100,000 users. Despite such success, these systems suffer from issues related to asymmetries in usage patterns such as skewed commute flows and preference for down-hill biking. In this talk, we provide an analysis of the shared biking system in Barcelona, Spain, called Bicing, which currently has 3,000 bikes and over 100,000 users.

Using three months of Bicing usage data along with external context features such as elevation, weather, and population density, we have constructed predictive models of station usage to predict bike availability. We present the results of our prediction algorithms and close by discussing their implications on station placement, design, and the potential for mobile applications.