*3.1. Data Collection*

Advances in (mobile) technology open up ways to improve the location and travel monitoring in the Active2Gether system.

First, for determining modes of transportation, we currently use daily questions in combination with location data (i.e., prompted user input). At the time of the design, this was a reasonable choice, considering the state-of-the-art and the consequences on battery consumption during full-time use of accelerometers in the mobile phone. Nowadays, the Google Activity Recognition API and the iOS CM Motion Activity class would be logical candidates, as they are the de-facto standards for providing location information [28,29]. Power consumption remains an important issue, however.

Second, we used a complicated questionnaire for reporting significant locations and their characteristics (such as relevant floors). It is difficult and cumbersome to answer, and also difficult to update during the intervention. Therefore, we recommend to automatically detect significant locations [11], which can also be done via Google location services. A remaining drawback, however, is that the users are required to provide this privacy-sensitive information.
