SAFEBIKE: A Bike-sharing Route Recommender with Availability Prediction and Safe Routing

Weisheng Zhong, Fanglan Chen, Kaiqun Fu

Abstract

This paper presents SAFEBIKE, a novel route recommendation system for bike-sharing service that utilizes station information to infer the number of available bikes in dock and recommend bike routes according to multiple factors such as distance and safety level. The system consists of a station level availability predictor that predicts bikes and docks amount at each station, and an efficient route recommendation service that considers safety and bike/dock availability factors. It targets users who are concerned about route safeness and station availability. We demonstrate the system by utilizing Citi Bike station availability and New York City crime data of Manhattan to show the effectiveness of our approach. Integrated with real-time station availability and historical crime data resources, our proposed system can effectively recommend an optimal bike route and improve the travel experience of bike users.