Abstract:The rapid growth of Internet based ride-sharing brings great changes to residents' travel and city traffic. However, few studies had employed empirical data to examine the unique travel patterns of internet based ride-sharing trips. In this paper, we compare taxi trip records and internet based ride-sharing trip records provided by DiDi company. Results reveal many interesting findings that had never been reported before. From the viewpoint of service patterns, ride-sharing mainly increases supplies in hot areas and peak hours. By applying a non-negative matrix factorization method, we find that ride-sharing principally serves as an approach for commuting. So, as an effective supplement to traditional taxi service, it regulates spatial and temporal supply-demand imbalance, especially during morning and evening rush periods. From the viewpoint of individual behavior patterns, we use a clustering method to identify two kinds of Internet based ride-sharing drivers. The first kind of drivers usually provides ride-sharing along daily homework commuting. Trips served by these drivers have relatively constant origin-designation (OD) pairs. The second kind of drivers does not serve regularly and roams around the city even in working hours. Therefore, there are no constant OD pairs in their ride-sharing trips. Counterintuitively, we find that home-work commuting drivers account for only a small part of total drivers and they only serve a small number of commuting trips. In addition, Internet based ride-sharing is not just traditional hitchhiking worked through mobile internet. We find that internet based ride-sharing drivers intend to make long distance trips, and they intend to detour further to pick up or drop off passengers than traditional hitchhike drivers since they are paid. All these findings are helpful for policy makers at all levels to make informed decisions about deployment of Internet based ride-sharing service. This paper also verifies that big data analytics is particularly useful and powerful in the analysis of ride-sharing and taxi service patterns.

Assessing the potential of ride-sharing using mobile and social data

A Shared Ride Matching Approach to Low-Carbon and Electrified Ridesplitting

The Potential of Ridesharing Adoption and its Effects on CO2 Emissions and Customer Experience

Dynamic ride sharing using traditional taxis and shared autonomous taxis: A case study of NYC

Ride Sharing and Dynamic Networks Analysis

Real-Time City-Scale Taxi Ridesharing

On-demand high-capacity ride-sharing via dynamic trip-vehicle assignment

Spatio-Temporal Mobility Patterns of On-demand Ride-hailing Service Users

An Empirical Study on Travel Patterns of Internet Based Ride-Sharing

Sharing Behavior in Ride-Hailing Trips: A Machine Learning Inference Approach

Scaling Law of Urban Ride Sharing

Can dynamic ride-sharing reduce traffic congestion?

Car sharing through the data analysis lens

Quantifying the benefits of vehicle pooling with shareability networks

Quantifying traffic emission reductions and traffic congestion alleviation from high-capacity ride-sharing

A User-Oriented Taxi Ridesharing System with Large-Scale Urban GPS Sensor Data

Adoption and frequency of use of ride-hailing services in a European city: The case of Madrid

Addressing the minimum fleet problem in on-demand urban mobility

OpenStreetCab: Exploiting Taxi Mobility Patterns in New York City to Reduce Commuter Costs

Spatiotemporal variability of ride-pooling potential -- half a year New York City experiment

Online Equitable Ride-sharing Car Distribution in a Congested Traffic