What is it about?

This research introduces a new, detailed dataset called LIFETRACE to help scientists study how people form long-term physical activity habits and how this affects their well-being. Unlike most existing datasets that focus on clinical populations or artificial settings, LIFETRACE focuses on general population, mainly university students, going about their daily lives for 14 weeks.

Featured Image

Why is it important?

LIFETRACE is uniquely important because it moves beyond the limitations of datasets focused on clinical populations by richly integrating objective activity data (from wearables), subjective daily mood and well-being assessments, and repeated physiological measurements (like blood lipids and BMI). This combination is critical for developing and benchmarking advanced machine learning models for personalized digital health interventions, especially given the randomized design that explores the causal impact of different environments (city vs. park walking) on behavioral and metabolic outcomes. The data’s comprehensive nature and focus on habit mechanisms make it essential for informing future research and public health strategies aimed at boosting long-term health adherence.

Perspectives

As one of the authors of LIFETRACE, I believe our greatest contribution lies in the intentional fusion of diverse data streams to study a single, cohesive human goal: forming a healthy habit. The real world is not a lab, and healthy habit formation is not purely a psychological or purely a behavioral phenomenon: it’s an intricate dance between what the body is doing (wearable steps, blood lipids), where the person is (city vs. park), and how they feel (daily questionnaires).

Francesco Bombassei De Bona
Universita della Svizzera Italiana

Read the Original

This page is a summary of: LIFETRACE: A Longitudinal Multimodal Dataset on Daily Physical Activity, Well-Being, and Habits, Proceedings of the ACM on Interactive Mobile Wearable and Ubiquitous Technologies, December 2025, ACM (Association for Computing Machinery),
DOI: 10.1145/3770676.
You can read the full text:

Read

Contributors

The following have contributed to this page