A new tool developed by Princeton researchers provides gig workers access to aggregate wage data that companies do not provide, a boon for rideshare labor advocates.
Platforms like Uber and Lyft use proprietary AI algorithms to determine pricing and payment for each ride. According to the researchers, although drivers have access to their own payment history, the platforms make accessing this data difficult or opaque and do not provide ways to aggregate the data at scale.
To make collecting this data easier, researchers at Princeton’s Center for Information Technology Policy developed an application, FairFare, which lets labor organizations automatically collect and store detailed wage and payment information from their members. FairFare allows researchers to construct a large-scale dataset that aggregates individual worker data across many thousands of rides to shed light on the working conditions of rideshare drivers.
“FairFare was developed to increase transparency into opaque algorithms that make decisions for people,” said Andrés Monroy-Hernández, associate professor of computer science and principal investigator on the research, which was published on October 15 in ACM Transactions on Computer-Human Interaction.
Monroy-Hernández and his team built the tool in collaboration with a Colorado labor union, Colorado Independent Drivers United. The goal is to give drivers and unions aggregated data that platforms already have access to, allowing rideshare drivers as a group to advocate for greater transparency about their wages and income.
Samantha Dalal, a postdoctoral researcher at Princeton and a lead author on the paper, worked with rideshare drivers in Colorado. “The platforms already have access to workers’ pay data,” she said. “Unions should similarly be empowered to have that same access.”
Drivers access FairFare as a web application on their phones. Once they consent to sharing data, FairFare can easily collect information from a driver’s rideshare account.
However, determining the “take rate,” the amount the platforms take from each ride, also requires knowing how much customers are charged. Uber does not routinely and transparently share customer pricing data with drivers.
To gather the available customer pricing data, researchers used a software interface tool made available by a third-party application that provides income and employment verification for rideshare workers. By pairing these two sources, the researchers were able to glean complete payment information for each transaction and aggregate that data into a broader snapshot.
With help from the labor union, the researchers recruited 45 drivers to use FairFare over the course of several years and gathered payment data on 76,625 rides. This payment data was then analyzed by researchers to determine the take rate.
The data yielded unexpected results, said Varun Rao, a doctoral student at Princeton and another lead researcher on the project. “The average take rate across thousands of rides was about 30%, which was lower than the drivers and their labor union initially thought,” he said. But while the take rate was lower on average, the data also revealed that it was highly variable. A few rides had take rates as high as 80%, with some as low as 10%.
“The fact that take rates exhibit such high variation means that drivers have a difficult time in establishing a stable work routine, because they can’t predict how much the platform will take,” said Dalal.
Examples of very high take rates helped make the case for the transparency required by Colorado’s Transportation Network Transparency Bill, which was signed into law on June 5, 2024. The new law requires companies like Uber and Lyft to share payment details after each trip, including what the customer paid and what the driver received.
The researchers’ hope, said Monroy-Hernández, is that companies will voluntarily share data, so FairFare isn’t necessary. Until this happens, FairFare will continue to be a useful tool for gig workers and the labor unions that advocate for them. Today, FairFare has collected data on four million rides from over one thousand accounts.
Even in states and cities where the government regulates rideshare platforms, the need for data is crucial, Monroy-Hernández said. For example, Seattle has a law that allows drivers to receive compensation if they have been wrongly deactivated by a platform. But that compensation is only granted if a driver can show what his daily earnings were for 12 weeks before the date of deactivation. Prior to FairFare, collecting this data was onerous for drivers and unions, said Rao. FairFare allows drivers to easily print out a detailed report of their payment history.
As more companies use opaque algorithms for tasks like hiring and surveillance and social media, the need for more transparency is growing, said Dalal. Research like FairFare, she said, can demonstrate “what needs to happen for transparency to actually translate into protections for consumers, workers and the public.”
The paper, “FairFare: A Tool of Crowdsourcing Rideshare Data to Empower Labor Organizers,” was published October 15, 2025 in ACM Transactions on Computer-Human Interaction. In addition to Monroy-Hernández, Dalal and Rao, co-authors include Dana Calacci of Penn State University; Catherine Di and Kok-Wei Pua of Princeton; Andrew Schwartz of Cornflower Labs and Danny Spitzberg of the University of California-Berkeley.
