“How well did I sleep last night? How relaxed am I right now? Will I feel better or worse tomorrow?” Personal informatics and self-tracking systems contribute to the expectation that personal health and wellness questions like these can be answered with data. Because visualizations play a pivotal role in many PI systems by making tracking data available to end users, design decisions related to visual encodings are deeply implicated in perceived associations between self-knowledge and pervasive personal data. This is particularly true for vulnerable populations like those who self-track to manage serious mental illnesses. This talk will focus on a recent co-design project conducted in collaboration with individuals diagnosed with bipolar disorder, a serious mental illness characterized by difficult to predict mood swings and often controlled through therapeutic self-tracking. This work provides a basis for a discussion of 1) sense-making challenges related to the representation and interpretation of personal data and 2) the benefits, risks, and limitations of participatory approaches to designing personal data visualizations that better reflect lived experiences.