Mart App

The problem

For his dissertation at Freie Universität Berlin, Roland Toth wanted to study how people use their smartphones. That meant combining four methods in one app: an initial survey, several short surveys a day, automatic collection of usage data on Android, and submission of usage data on iOS.

Existing tools each covered only part of this, required programming skills, or collected extra sensitive data such as GPS locations. Using several apps at once would have been an extra burden for participants. So MART was developed as a custom app.

The solution

I built the first version of MART in 2022 and 2023 as a developer at nvii-media, for Android and iOS in parallel. In 2025, working as a freelancer for the Weizenbaum Institute, I completely revised the app so that other researchers can use it for their studies without programming skills. For development I set up a test back end with Directus.

The main features:

• Login with participant ID and password
• Welcome page and introduction to the study
• Initial survey and recurring short surveys
• Many question types that can be combined freely: text field, number field, single and multiple choice, scales and sliders, timed questions, audio and photo recordings
• Automatic collection of usage data on Android
• Entry of usage data on iOS through a form
• Submission of device information
• Surveys and pages are configured in the back end
• Dark mode and light mode

Technical challenges

A custom Capacitor plugin for Android

Capacitor connects a web app to the native features of Android and iOS. There was no suitable plugin for reading usage data, so I wrote one in native Android code. It requests the required permission, reads the operating system's event logs and passes the data to the app. From the second collection on, only new events are transferred, with a small overlap so that no data is lost if something goes wrong.

Usage data on iOS

Apple does not let apps access the usage data of other apps. On iOS, MART therefore guides participants to the Screen Time overview on their iPhone with an illustrated walkthrough, and they enter the values in a form.

Surveys that appear on their own

Short surveys are shown automatically as soon as defined conditions are met, such as a time window in a participant's personal schedule. A notification points them to it.

Question types that combine freely

Studies differ a lot. That is why the question types can be combined freely and configured in the back end, without publishing a new version of the app.

Privacy

Participants log in with a participant ID instead of personal data. Only what the study needs is collected.

The result

• With the first version, 1,780 participants with Android and iOS devices took part in the FU Berlin study for one week.
• According to the dissertation, this produced one of the most comprehensive data sets on mobile media use currently available.
• The app was presented in the journal Publizistik and at the International Conference on Computational Social Science 2023 in Copenhagen.
• The study results are published in the dissertation Rethinking Mobile Media Use by Roland Toth (FU Berlin, 2026).