Examined if it was possible to predict the emoji label using phone usage data so that targeted interventions could be made.
G-Moji is a mental health self-help app that promotes digital wellbeing for youngsters, funded by many healthcare institutions and spearheaded by Garage 2020.
The pilot study for this app measured diverse data from phone-usage such as movement, app-usage, and light intensity of their surroundings. It also polled users daily to describe their mood for the day using one of 14 emojis. The initial goal of the project was to see if it is possible to predict the emoji label using the phone usage data so that targeted interventions could be made. After prototyping and iterating we concluded that there was not much gain from using ML on phone usage data.