Studying cell-to-cell heterogeneity is essential to understand how unicellular organisms respond to stresses. We introduce a single-cell framework that enables the study of interactions between photosynthetic traits within individuals with the same genotype and cellular context, along with common histories. Our approach combines single-cell imaging of chlorophyll a fluorescence with machine learning and we study light stress responses in Chlamydomonas reinhardtii as a proof-of-concept. This framework allows us to score the extent of high-light responses such as state transitions (qT) and high-energy quenching (qE), and reveals significant cell-to-cell response heterogeneity. We uncover a strong correlation between qT and qE at the individual level, an interaction that cannot be detected by bulk measurements. Our study highlights the value of single-cell phenotypic analysis for uncovering relationships between traits. We detail the key aspects that come into play to generalize the method to other multi-trait systems.