CloudTrip: Exploring physiological data-enabled multisensory metaphorical narratives for everyday stress management

In modern society, an increasing number of people are suffering from the harmful effects of chronic stress, while data-enabled reflection has been shown to be an effective approach to stress management. However, most existing human–data interaction designs primarily focus on retrospective data review, often failing to foster engaging and meaningful reflection. In this paper, we present CloudTrip, an immersive application that represents the users’ data throughout a day as an experiential and interactive digital story for mental health self-reflection. Specifically, CloudTrip metaphorically conveys users’ daily stress-related physiological data into narrative experiences by dynamically mapping the data onto changes in visual and auditory elements of characters and virtual environments. Users can record reflective content during the experience and receive responsive feedback generated by a large language model, thereby forming an interactive narrative loop that further deepens reflective engagement. We conducted a 14-day within-subject study with 20 participants, comparing CloudTrip with a traditional line-chart visualization using a mixed-methods evaluation to assess its impact on the reflection experience and reflection support. Results show that CloudTrip was associated with a cardiac autonomic pattern compatible with higher arousal, increases in self-reported valence, arousal, and dominance, more favorable motivational ratings for the experience, stronger perceived support for reflection, and promise for repeated use. Finally, we discuss design implications for integrating data-enabled multisensory storytelling techniques into human–data interaction systems to promote everyday reflective behavior for mental health management.