What's new on Beiwe

Digital phenotyping research platform

Coming Soon
April 30, 2021

Foreground Service and Ambient Audio Collection Release

We are currently testing and preparing to release our foreground service and ambient audio collection for Android. These features will be available starting in version 3.1.0 of the Android Beiwe app, which will be available from studies.beiwe.org/download and the Google Play Store in the coming weeks. Please see below for additional details.

Foreground Service: the Android app now has a persistent notification that tells you it's collecting data. This should accomplish three things:

  • Re-enable data collection from the accelerometer and gyroscope when the app is in the background, for Android OS versions 9 and above.
  • Enable us to push updates to the Google Play Store. Google enacted a policy last month prohibiting apps that use the background location data permission unless Google specifically approves the app. This policy hasn't affected existing versions of the Beiwe app on Google Play, but it does block us from updating the app until we make the change. So the Foreground Service enables us to keep collecting location data without doing it from the background.
  • Enable ambient audio collection.

Ambient Audio Collection: this is a feature that you can toggle on or off for a study on the Device Settings Page. It's off by default for all current and new studies. Here's how it works:

  • The app records data to an unencrypted temporary file on the phone. Every 15 minutes, it writes the temp file to a permanent encrypted file, deletes the unencrypted temp file, and then starts over again. Only the encrypted files get uploaded. So there is temporarily unencrypted data on the phone, but it gets purged every 15 minutes. This is the same architecture Beiwe has been using for the Audio Surveys since 2014. Eli devised an architecture for the text files that lets the Beiwe app write them so they're encrypted line-by-line, but we haven't been able to come up with an easy parallel solution for the audio files, so we use this temp file paradigm instead.
  • It records all data as MP4 files, using the same parameters as the regular/compressed audio surveys.
  • These audio files are not batched into hour-long chunks the way most of the passive data files are. So when a researcher downloads them, they'll get a series of MP4 files that should each be about 15 minutes long. The filename includes the timestamp from the end of the recording, not the beginning.

Please check back for updates and if you have any questions don't hesitate to reach out to kcarlson@hsph.harvard.edu.

- Beiwe team

Create your own newsfeed