Over the last few years, the use of smartphones has significantly changed our relationship with technology. Their connectivity, mobility and versatility have made them an essential part of our everyday life. Moreover, through their sensors, they are able to record suitable data to study users' emotions. Therefore, affective computing is increasingly focusing on these devices as objects of investigation.However, studying emotions remains a complex task. Indeed, inferring a person's emotions involves considering many features, creating specific tools and designing an adapted methodology. Experimenting on smartphones brings many challenges in order to record usable data. Among other emotional components, the subjective feeling, which could convey a verbalization of the emotional experience, needs specific tools to be assessed. These tools should be adapted to the specificity of the experimentation context and help the users report their feelings.In the scope of a project on automatic emotion recognition through smartphone sensors, we developed an android application to assess users' subjective feelings: Geneva Emotion Wheel Mobile Interfaces (GEWMI). It is based on an existing emotional measure instrument: the Geneva Emotion Wheel (GEW), that we adapted to mobile devices. This digital instrument was developed as a standalone and configurable open source application for android devices. We tested this application with two experiments with two distinct emotion elicitation protocol designs. The results show the flexibility of this tool to perform analysis on different levels.