Borapp Buzz is an experimental app which is part of an ongoing scientific research project on trying to study the relationship between phone usage and emotions.
Following a novel, smart, patent-pending algorithm, the app will create a notification that suggests you to visit a website – currently you can enter a URL of your choice – when it thinks that you might be bored. It does so by trying to recognize usage patterns that typically occur when people are bored. To work well, it relies on the assumption that people exhibit similar phone usage patterns when they are bored – which may or may not be the case for each and every phone user.
Currently, we are testing different parameters of the boredom-estimation algorithm. You already can download Borapp Buzz from Google Play for free and be part of that process.
Borapp Buzz is maintained by researchers from the scientific group of Telefonica R&D, Barcelona, Spain:
- Martin Pielot, associate researcher at Telefonica Research
- Nuria Oliver, scientific director at Telefonica Research
How does the boredom detection work?
We describe the research behind Borapp in this blog post. In summary, we had people frequently indicate how bored they feel through an app. At the same time, the app was collecting usage patterns. We then used machine-learning technology to train algorithms that recognize phone-usage patterns that typically co-occur with boredom. If you wish to know more, you can read the scientific paper, which discloses all the interesting details.
What data does Borapp collect?
The data that the app currently collects is for each time it suggests content via a notification:
- (1) whether you click it,
- (2) if yes, after how much time and how much time you spent in the browser, and
- (3) the status of the boredom prediction when the notification was fired.
This data is anonymous: it won’t allow to identify, locate, or contact you.
Why does Borapp need access to accessibility events, notification events, and location?
The experimental boredom-detection algorithm takes signals from these three sources into account when estimating whether you are bored. For example,
- the accessibility services allows to see the package name of the app, which is currently in foreground (e.g. com.whatsapp) — the accessibility service allows to implement this in the most battery-preserving way;
- the notification listener tells the algorithm, when the last notification was received (e.g. 10:45am, ‘com.whatsapp’) and when notifications were dismissed;
- the location data is used to estimate whether the you are at home, at work, or elsewhere.
All these signals will be processed locally on your phone and never be communicated to our servers!