FamApp

FamApp (Familiarity Application) (previously, SoLin) is a project of TU Darmstadt. Help us create a model to automatically estimate your familiarity with your contacts!

In order to create the model, communication data is collected from the call and SMS logs as well as instant messaging (IM) applications – WhatsApp and Threema – per selected contact. Furthermore, the application also collects logical location information (e.g. home, work, etc.). Subsequently, from these communication data, the application extracts certain so-called indicators (e.g. number of calls on Sundays, number of emoticons in IM messages, number of messages sent at work, etc.).

Your role as a user is to answer a questionnaire for each selected contact (User Assessment), and also inform the application about your logical location (Note: We save the corresponding cell tower/WiFi-hotspot information for each logical location – no GPS-coordinates are stored).

Our model shall be created with the help of the collected communication indicators as well as the answers to the questionnaires. Through this model, it should be possible to estimate the type and strength (familiarity) of a user’s relationship with his/her contacts. For example, it should be possible to determine if a contact is a (close/distant) colleague, friend, or a family member.

How could you profit from this? Here are a few reasons why you should support this research work:
- Automatic decision support: With the help of this information, the smartphone could decide if and when it puts through or blocks notifications. For example, your phone can identify your boss and automatically block his calls past working hours.
- Preserve your privacy: You could share content on online social networks with people whom you really want to address. Positive side-effect: You also receive fewer messages from people whom you are not close to.

For your participation, you can win Amazon vouchers worth up to 50€!

Privacy statement:
During the user study, the application will access the contact list of the user to prepare a questionnaire in plain-text so as to differentiate between the respective contacts. Furthermore, the application requires access rights to the usage statistics as well as incoming notification, so that the above-mentioned indicators can be extracted. The results from the user questionnaires will be saved on the user’s smartphone, together with the extracted indicators of the respective contacts.

The collected dataset will be uploaded on to a server, where they will be combined with datasets collected from other users, and then analyzed using machine learning algorithms. They shall be treated completely anonymously, such that no personal information escapes the smartphone. Each dataset will be appended with the shortened device ID for each user. Therefore, no data will be saved on the server that allow for any inference of the respective user identities. The collected data will be exclusively used for research purposes. The release of the personal data to unauthorized third parties is ruled out.