GLOBEM Dataset

The GLOBEM dataset is a multi-year, publicly available collection of passive sensor data from smartphones and wearables, coupled with wellbeing metrics from hundreds of users. Research using GLOBEM focuses on developing and evaluating algorithms for longitudinal behavior modeling, particularly in areas like depression detection, often employing machine learning techniques including large language models and domain adaptation methods to improve cross-dataset generalizability. This resource is significant for advancing the field of personalized health monitoring and behavior prediction by providing a standardized benchmark for evaluating the robustness and generalizability of algorithms across diverse populations and time periods.