Toward an ecology of emotion in everyday life

Human emotional life is rich and varied, yet psychological science often studies it in decontextualized ways, asking participants to rate feelings without capturing the situations that evoke them. Here, we map the ecology of everyday emotion by analyzing the link between affective experience and daily events. We collected thousands of real-time verbal descriptions of momentary affect from participants over 70 days. Using transformer-based language models, we identified granular topics corresponding to daily situations, from 'morning routines' to 'socializing with friends.' Our analysis revealed that the mapping between situations and emotions is not one-to-one, but many-to-many. For example, a single situation, such as 'reading books,' was associated with a wide spectrum of emotions, while a single emotion word, such as “angry,” appeared across a diverse range of situations. This data-driven approach moves beyond simple emotion labels to build a foundational ecology of human affective experience.