If I Tap It, Will They Come? An Introductory Analysis of Fairness in a Large-Scale Ride Hailing Dataset: An Abstract

Ride hailing service market is by far the fastest growing industry. Increased consumer demand resulted in a significant shift from traditional taxis to ride hailing services (Pyzyk 2019). According to a 2017 report of Goldman Sachs, this industry is expected to reach a market size of $285 billion by 2030 (Huston 2017). Uber is the largest ride hailing company in the market, followed by Lyft and few other relatively small companies such as Via, Gett and Juno (Lam and Liu 2017). There are limited regulations on these ride hailing services which are black-box algorithmic decision makers. Consequently, there are growing concerns about algorithmic fairness in ride hailing.

To investigate fairness in ride hailing services, we analyzed data provided by the City of Chicago on 73,247,231 ride hailing Uber, Lyft, and Via trips combined between November 2018 and June 2019 for 5,459,609 drivers. City of Chicago has been first to publish city level data on ride hailing trips in 2019. Our findings indicate that there are some concerns in terms of fairness in their practices. As our findings suggest, low-income neighborhoods pay higher prices for their trips than high income neighborhoods. Additionally, consumers from minority and low-income neighborhoods have less ride hailing service usage as opposed to consumers from white dominant, high-income neighborhoods. Our findings also show that popular pick-up and drop-off neighborhoods have higher ride hailing prices which explains the demand-based surge pricing practices of these ride hailing companies. Finally, we found that young and active consumers that use ride hailing services in their communities pay higher prices in comparison to other populations of different ages.

This paper contributes to an increased understanding of fairness in ride hailing services by analyzing the detailed large-scale ride hailing dataset obtained from Chicago Data portal. The results provide a preliminary understanding about the demographic patterns of consumers’ ride hailing usage. This is one of the first city level datasets on ride hailing services that is publicly available and provides insights about the practices of these services. This research contributes to the literature of ride hailing services as well as practices of these services in terms of fairness. Managerially, by better understanding the factors that create unfair practices in the ride hailing market, marketers, researchers and policy makers can offer solutions or work together to set regulations aiming to prevent disparate impact in the ride hailing industry. Finally, our findings suggest that further exploration of ride hailing services with sophisticated machine learning techniques can provide insights as to how fair ride hailing services are, how they affect consumers and to what extent these services contribute to the growth of communities.