Breathless is a live audiovisual performance designed using biosensor data and audio recordings of breathing rhythms collected from myself while watching video documentation of protests in Iran over a 6-month period. Statistical analysis and machine learning were the techniques consecutively implemented to process biosignals and audio recordings. The goal for such processing is to make sense of the time-series data towards an acoustic performance. While making use of both libraries of data, the performance features live-streamed sensor data captured while watching the video documentation of the protests as a re-enactment of the data collection phase. As a critique of the field of affective computing that aims to automate emotions as well as data-driven media art practices that make unquestionable use of data processing techniques, Breathless intends to open up the grounds for alternative treatments of data towards critical possibilities. Through leveraging the atmospheric qualities of sound and the experience building capacity of the live performance, the goal is, thus, to opens up a space for relationality and empathy building.