Join in now!
You can take part at any time
Together for safer cycling in Mannheim
A project by QuadRadEntscheid Mannheim and BlossomEdge
Motivation
Who we are and why we're doing this
How many dangerous moments do you experience on your bike that never show up anywhere? No accident, no statistic, no political pressure. Today, cities respond almost exclusively to accident data — everything that ends without injury stays invisible.
QuadRadEntscheid Mannheim and BlossomEdge want to change that. BlossomEdge has developed an app that allows cyclists to easily capture and document dangerous situations while riding. Together, we are launching a project in Mannheim to show — for the first time — where it is truly dangerous for cyclists in this city, without waiting for the next accident. The results and anonymised data will be made available to QuadRadEntscheid Mannheim and will feed directly into their work: as a data foundation for concrete demands to the city.
Technology
How does it work?
Go-Bike is an app by BlossomEdge for iPhone and Android. You mount your smartphone on the handlebar, start the app — and ride as usual.
The core idea is simple: if you experience a dangerous situation during your ride — a risky overtaking manoeuvre, a confusing intersection, too little space, or any moment where you simply feel unsafe — you tap briefly on the screen. Go-Bike automatically saves the 5 seconds before and 2 seconds after that moment: with location, timestamp, and a short video clip. This creates a precise, geolocated record of exactly the situations that would otherwise go undocumented.
In addition, Go-Bike analyses the camera feed in real time using AI and can automatically detect potential risk situations and alert you. This feature is in active development — the brief tap on the screen remains the most reliable way to capture a situation.
All recorded rides together form an aggregated picture: where do dangerous situations cluster? Which intersections, which streets keep appearing? That is the basis for the heatmap that will be handed over to QuadRadEntscheid Mannheim as part of the project.
Join
How it works
Joining is straightforward. You need:
- An iPhone or Android smartphone
- A handlebar mount for your smartphone
- Your normal everyday routes
Download the Go-Bike app
Mount
Mount your smartphone on the handlebar using the mount — camera facing forward.
Start the app and ride
Just ride as usual — the AI automatically detects and documents risks.
Quick tap when it gets dangerous
Want to manually flag a dangerous spot or moment? A brief tap on the screen is all it takes.
If you don't have a mount yet: we recommend this universal handlebar mount, which works with all smartphone sizes — available on Amazon for around €15.
Handlebar mount on Amazon →Launch giveaway
Anyone who uploads at least 20 rides of over 2 km each by 31 October will be entered into the draw. QuadRadEntscheid is giving away 3 QuadRadEntscheid Big Bell bike bells — one per winner. To take part, simply contact their team by email at hallo@quadradentscheid.de.
You can join at any time — the project runs on a rolling basis. Every recorded ride helps! Please also invite friends and family: the more rides are captured, the more meaningful the picture of Mannheim becomes.
Your Impact
What happens with the results?All recorded and shared rides feed into an aggregated analysis. The data is anonymised in the process — see for more details.
The central output is a heatmap of Mannheim: it shows where risk situations cluster — the most dangerous spots in the city, made visible through your rides. BlossomEdge will provide this heatmap to QuadRadEntscheid Mannheim, who will use it in their work with the city: as a data foundation for concrete demands, as evidence in political conversations, and as a basis for targeted infrastructure improvements.
During the project, we will keep you regularly updated on interim results — through QuadRadEntscheid Mannheim’s channels. You will see how the picture of Mannheim sharpens with every additional ride.
Privacy
We are aware that an app that records during a ride raises questions about data privacy.
Which data is transmitted and when
During the ride, the AI on your smartphone analyses the camera image and the sensor data. This happens locally and without an internet connection. A continuous video of the ride is neither stored nor transmitted. Stored on your smartphone are:
- the GPS track of your ride with timestamps
- inertial data from your smartphone's sensors (accelerometer, gyroscope, magnetometer)
- a short video clip when the AI detects a risk situation or you tap the screen
- the model results of the detection
After each ride, a short summary is being transmitted (start and end time, number of detected situations). The ride data itself is only transmitted when you upload a ride or have switched on automatic upload. The transmission is encrypted. The servers are hosted by Amazon Web Services (AWS, region: eu-west-2). There, further AI models, including language models, analyse the video clips. They recognise relevant features of a scene (e.g. tram tracks) and classify what happened in a scene (e.g. dooring, right of way not given) and which features of the infrastructure contribute to the risk (e.g. poor visibility). We also use the results of this analysis to further develop our models.
What we do concretely
You decide what is shared. Recorded rides are not automatically transmitted unless you explicitly activate this feature. After each ride, you decide whether and which data you want to contribute to the project. Only data you actively share feeds into the analysis.
Video recordings are anonymised before sharing. Faces and licence plates are automatically blurred on our servers after upload, before any video data is shared with third parties. No original recordings are passed on.
Rides are anonymised. The first and last sections of each ride are automatically trimmed. The start and end of your ride — typically your home or workplace — are therefore not part of the analysis.
No personal data. No personal information is linked to or shared with rides — no email address, no name, no user profile.
Who receives the data?
The anonymised analyses are shared with QuadRadEntscheid Mannheim, who uses them in their work to improve cycling infrastructure in Mannheim. What always applies: shared data is processed in such a way that it cannot be used to identify you as an individual. Your identity remains protected.
What is shared even without your explicit consent?
Even without your opt-in, BlossomEdge may share fully aggregated or anonymised data from all rides — for example, heatmaps showing where many cyclists experience risks, or average risk scores per road segment. This data cannot be attributed to any individual person or ride. This is covered by the Terms of Service you agree to when creating an account. If you additionally give your opt-in, your individual rides (anonymised, but not aggregated with other rides) may also be shared.
Anonymous identifier
If you consent to data sharing, your rides are assigned a random identifier so they can be analysed together (e.g. to detect patterns across multiple rides). This identifier is not linked to your name, email address, or account.
Full control, at any time
You can delete all your data — locally on the device and in the cloud — at any time directly through the app. You can revoke your data sharing consent at any time under Settings > Data Sharing. Data that has already been shared cannot be recalled, as it has been anonymised and can no longer be linked to you.
Participation in the project is voluntary and can be ended at any time.
Questions about privacy? support@blossomedge.ai
Frequently Asked Questions
Make Mannheim safer — one ride at a time.
Free. For iOS and Android.