Using Data to Optimize Resources
How UIP uses prediction models to offer a more efficient and user-friendly bike share experience
Oslo City Bike, UIP’s flagship bike share system, produces an average of 9.8 trips per bike per day. This critical number, which shows how well the system optimizes its resources, makes Oslo’s bike sharing platform one of the most efficient in the world. While other systems focus on their number of bikes, UIP knows from experience that a lot of hardware doesn’t necessarily equal a smart and successful system. Unused bikes standing around the city simply take up sidewalk space without contributing to what should be the goal of bike sharing — making urban mobility more efficient.
Since its launch in 2016, Oslo’s new system has become so popular, and gets so much daily use that the system is growing quickly to keep up with the demand. New stations and bikes are being added all the time, but there’s still the challenge faced by all well-used bike share systems: users being met with too many empty and full stations, especially during the daily rush hours.
To address the issue, UIP teamed up with leading design firm Designit to tackle this challenge. The collaboration was part of the Design-Driven Innovation Program (DIP), a program in Oslo that finds creative and innovative ways to improve products and services for Norwegian businesses.
“It is always challenging, but very fun, to work on improving services that customers really like,” says Designit’s senior product designer Siri Yran, who, together with senior designer Emilie Olsen, worked closely on this collaboration. “The challenge is understanding why the service works so well, and to know how to maintain this as the service changes and evolves. It is rare to work with a client that is so concerned with the needs and desires of their users as UIP is.”
To address this primary user need— having fewer empty and full stations in the mornings — was one of the first concepts created from the “DIP” initiative. Called “Morgenfugler” (“Morning Birds”), the idea is to analyze historical system data and determine which stations are the first to become full and empty during the morning rush, therefore negatively affecting the experience for commuters. The solution was to place Oslo City Bike team members at some of these stations during the busiest time of the morning. These workers are there to take bikes out of the stations as they get filled up, improving the commuter experience by allowing more users to park their bike where they want.
In addition, UIP placed team members at the stations that get emptied most quickly in the morning. Here, the stations are continually filled with fresh bikes from a stock that was brought to the station and secured nearby the night before. In this way, the system becomes available for more users, and improves the efficacy of the whole system.
To achieve the maximum effect from this initiative for both the individual user and the system as a whole, UIP utilized its historical data to determine exactly where it was best for the “morning birds” to be placed.
“After operating the city bike system in Oslo for more than 1.5 years, we have collected large amount of data that have given us a good understanding of mobility in Oslo. So we’ve tried to be smart in choosing the locations of the morning bird stations by basing it on our prediction models. If done correctly, the morning birds can improve the overall well-functioning of the system,” says Hans Martin Espegren, a data analyst at UIP. “Having morning birds located at various stations lets us collect additional data that will be analyzed to further improve the operation of our system.”
Since the project began in May, the data that Espegren has gathered shows a great effect already. On an average day, around 200 bikes are picked up from the popular downtown locks, meaning 200 more people can return their bike where they want each morning. The workers feeding new bikes into the system distribute an average of 45 additional bikes each day during the commuter rush at a designated station. In this way, the initiative helps to solve an immediate issue, but it’s also gathering data to improve the system for future development in Oslo and other cities. The workers take records of how many bikes they added or removed from the stations, thereby giving UIP’s data analysts valuable information about how many bikes and locks are needed to improve the system, and where they should placed.
Liisa Andersson, UIP’s COO, says that by using prediction models and user insight, UIP can better understand how to adjust its operation methods to guide the flow of the system in smarter ways. The use of data allows for predictions of when and where the system has the highest demand, and which stations have the greatest effect on the movement of the entire system.
“It’s about meeting user expectations,” Andersson says. “We’ve learned that our service needs to be available for users five minutes before they even know that they want to use it. The question is, how do we ensure that the bike is there for one user when they want it, and then available again for the next user in their five-minute window? How many different people can we connect with one bike? By using data, we can see exactly how people move, and how each trip connects to other trips. Having that knowledge helps us make the whole system better.”
UIP wants bike sharing to be a dynamic element in the movement of the city. One of the best ways to achieve this is by continually analyzing how the system works, and how information about user patterns and behavior can be applied to optimize every element of the system. As UIP expands to more cities, these types of projects will help UIP’s bike share systems to be as utilized as possible. By combining data with creative insights offered by designers, Oslo City Bike has shown how it can connect with customers’ needs at the street level to ensure an optimal balance between collective efficiency and individual needs.
