Sebastian Rühmann

Sebastian Rühmann

Computer Science
Universität Hamburg
First chapter

Project Available-Citybikes: Concept and Funding

Since 2014 I am a user of StadtRAD in Hamburg.

StadtRAD offers a shared bikes service in Hamburg. Customers lend bikes for 5 euro a year on 250 stations across Hamburg. The use is free for rides up to 30 minutes (otherwise 0,08-0,10 Euro/ min). Stadtrad holds over 3100 bikes in its fleet.

I use StadtRAD for urban travel basically everywhere. To commute from home to university, to visit friends and also for spontaneous trips. So why do I believe shared bike concepts like StadtRAD are great and everybody should use them?
- Shared bikes make transport simple, spontaneous, sustainable, affordable and fast(er).

The StadtRAD App allows users to simply and spontaneously unlock bikes and hop on for a ride. But shared bikes like StadtRAD are also a good addition to general public transportation. Almost every subway station in Hamburg has a adjacent StadtRAD station to lend bikes. Therefore shared bikes also solve the last mile problem.
In many cities urban planning has changed. While for decades driving time of car transportation has been prioritized, nowadays city planners all around the world are promoting cycling as an efficient means of transport. This leads to implementations of better bicycle lanes, bicycle "superhighways" and parking lots specific to bikes. Already today many people can travel faster if they switch from car to bike. (plus you don't waste time to find a parking)
Of course when we talk about transportation we don't come around talking about sustainability and shared bikes are doing really good compared to basically every other type of mobility.

So hands down I have to consider myself being a shared bike power-user.

The moment that let me reconsider

Like everyday I was going to rent a bike close to my home and commute to university in the morning. It was a sunny day, and I was going to accompany a friend on the way. She was new in Hamburg and I honestly bragged about StadtRAD. But this morning turned out to be different:

Usually the station closest to my place holds multiple available bikes. Not this day! We decided to check the App for nearby stations and indeed one station had two available bikes. As reservations are not possible, we rushed and were devastated as we arrived at another empty station.
I realized that I will be late for class and honestly reconsidered my personal use of shared bikes.

What I learned from this & Why reliability is important

The reliability of shared bikes is based on stations with available bikes. If I'm repeatedly not finding a bike at a nearby station I'm tending to doubt the reliability of shared bikes and switch to a different type of mobility. In fact not only the reliability but also the doubts are disadvantageous to build habits.
We use habits to simplify our life. When we are establishing habits we change from a decision-mode to autopilot. We rely on experiences that we take for granted to occur again. E.g. most people take the same way to work or university every morning and many people also use the same type of mobility.
Therefore when we doubt the reliability of shared bikes we won't establish a habit to use shared bikes for commute and instead use another type to achieve our destination.

The systemical reliability problem
To understand the reliability problem it is important to understand the source of the problem. The demand. The demand is underlying various factors like commute (rush hours), events, weather conditions etc. According to these factors the local availability is highly volatile.

How to solve the reliability problem

  • Reservations
    The concept of reservations is well-established along sharing providers. Emmy, a shared motor scooter provider allows users to reserve a unit up to 15 minutes before the rental period begins. Other to mention are scooter companies like Tier, Volt and so on. StadtRAD also offers reservations for cargo bikes, but not for regular bikes. I am unsure why they chose to do so. (I contacted the customer support and hope to report about it in the next article)

  • Endless bikes / perfect distribution
    In a dream world scenario we have an equal amount of demanded and available shared bikes. A real bike glitch! Unfortunately the efforts to allow teleportation have not been successful yet.

    So what about perfect distribution?

    Stadtrad has employed staff which repairs and distributes bikes to improve the availability. Anyhow, distributing is a demanding task and requires a person to load a van with bikes, transport them to more frequented stations and unload them. While this can adjust smaller availability differences between stations it cannot solve the systemical availability problem.

  • Prediction
    The truth to reliability of shared bikes is to accept that availability is not granted, but predictable. - In the mobility concept of a modern city shared bikes are one important module but not the only one.

    I'm coming back to my reconsideration moment: What if I would've known the night before that I cannot expect an available shared bike at the nearby stations. I'd have set my alarm earlier in order to take a bike at a further distanced station, probably I'd have chosen to take a different type of mobility, but most definitely I'd have not developed doubts about the reliability of shared bikes.

How we plan to implement prediction

When we are talking about prediction in technical terms we talk about artificial intelligence or to be more precise about neural networks. A neural network takes input from different sources and calculates combined likelihoods for an event to happen.

We want to answer the question of how likely it is to find an available bike at a local shared bike station. As described in the “systemical reliability problem” the availability is underlying the demand. Therefore our goal is to implement a neural network to process the most relevant factors for the local demand

The Project and funding

Mila and I pitched our idea to a jury of the Data literacy lab of the university Hamburg. Along with 9 other projects we achieved a funding of 10.000 Euros and support by the Data Literacy Lab. Over the next 12 months we are going to work on a solution for the systemical reliability problem by prediction.

We are going to conduct and evaluate a qualitative interview with an employee of StadtRAD. We are curious to study artificial intelligence in university courses, books and conferences. We are looking forward to making an impact with a prediction service for shared bike users and providers.
Ultimately we think the choice of mobility type should not be a question of availability but individual need. We are thrilled for the next months!

Sebastian Rühmann - Förderung Universität Hamburg
Data-Science Slam, University Hamburg

What can you expect here during the next months?

A monthly article with updates about the progress of our project and recommended ressources of artificial intelligence studies.

I'm going to explain method and tools we used, give insights in the academical collaboration at the Lab and illustrate pitfalls and workarounds along the journey.

The blog articles include the recommended ressources of artificial intelligence studies. I'm going to illustrate the idea of neural networks, refer to my favorite sources of learning (Books, online-courses, conferences etc.) and explain how we implemented our neural network. (also for non-developers)

Receive Updates
Leave your phone number for a notification when a new article released.