The implementation of a real-time algorithm to measure the engagement of the audience during a performance

Electrodermal activity (EDA) is a physiological signal which has been considered as one of the best indicators of sympathetic arousal, since it provides an indication of sweat gland activity. For this reason, it can be used to measure the engagement of the audience during a performance. This work exploits the use of wearable EDA sensors to measure the audience engagement during a live performance, testing a novel method that works in real-time. In particular, we collected data from 40 subjects participating in a jazz concert and used machine-learning based approaches to automatically identify the level of engagement during the show, minute by minute. Finally, we evaluate our algorithm using annotations reported during the live event. Our results suggest that improvements are needed in order to predict the audience's engagement accurately. However, this paper provides new insights originated from the particular experimental setting: a live music