Skip to content


Generalization of LOngitudinal BEhavior Modeling

GLOBEM is a platform to accelerate cross-dataset generalization research in the longitudinal behavior modeling domain.

GLOBEM aims to achieve the following goals:

  • 📦 Release the first multi-year longitudinal passive sensing dataset with ~500 unique participants
  • 🧰 Implement a series of longitudinal behavior modeling algorithms
  • 🔍 Prepare a series of cross-dataset generalization evaluation tasks
  • 🏅 Present the algorithms' generalizability benchmark on the GLOBEM datasets
  • 🔋 Enable researchers and developers to develop new algorithms with easy-to-extend templates


Datasets →

Multi-year passive mobile sensing datasets & well-being metrics.

Platform →

Generalizability evaluation platform and benchmark.

Publications →

Relevant publications based on the GLOBEM datasets and platform.

Team →

UW-EXP Team & Contributors. Contact us for any questions.