Thinking Outside the Data Box: Investigating the Potential of Data Manipulation for Self-Reflection on Personal Data

In the practice of personal informatics (PI), self-reflection is crucial for enhancing self-knowledge and driving behavior change. Numerous studies have focused on effectively interpreting and representing data to support self-reflection. However, despite their efforts, some self-trackers find themselves stuck in repetitive insights and stagnant process. For them, a fundamental shift beyond re-representing existing data could provide a significant opportunity. We explore data manipulation—altering the data value or structure—as an alternative approach. We conducted an exploratory workshop and a one-week field trial with 10 self-trackers, using five types of data manipulation. We found that data manipulation could revitalize self-reflection, uncovering diverse perspectives and overlooked aspects. It also fostered positive illusions and emotions, potentially setting the stage for behavioral change and engagement. However, it introduces perceptual distortions and has limited applicability, highlighting the importance of balanced use. We further discuss design implications for integrating data manipulation into future PI systems.