[Data] Unveiling Hidden Connections: Using Geospatial Analysis and LSTM Model to Foster Social Interactions in University Campus

This research investigates the capacity of geospatial data and machine learning techniques to predict and enhance social interactions in a post-COVID-19 university campus setting.

Authors

1. Haorui Zhou (corresponding author)

   • Department of Geography, University of California - Santa Barbara, CA, USA
   • 📧 h_zhou@ucsb.edu
   • 🍃 ORCID

2. Lauren Barley

   • Department of Geography, University of California - Santa Barbara, CA, USA

3. Harry Hebeler

   • Department of Geography, University of California - Santa Barbara, CA, USA

4. Thomas Shaner

   • Department of Geography, University of California - Santa Barbara, CA, USA

5. Somayeh Dodge

   • Department of Geography, University of California - Santa Barbara, CA, USA
   • 📧 sdodge@ucsb.edu
   • 🍃 ORCID

Funding

This student paper is supported by the U.S. National Science Foundation (Award BCS # 2043202).

Acknowledgements

S. Dodge gratefully acknowledges the grant from the U.S. National Science Foundation (Award BCS # 2043202) for supporting the student paper competition.

Abstract

This study investigates the capacity of geospatial data and machine learning techniques, specifically Long Short-Term Memory (LSTM) models, to predict interpersonal encounters and potentially enhance social interactions in a post-COVID-19 setting.

Our analysis focuses on a case study at the University of California, Santa Barbara (UCSB), where mobile location data from a group of individuals was used to train an LSTM model. The model utilizes a Hexagonal hierarchical geospatial indexing (H3) to detect encounters, achieving an accuracy of 0.85 in predicting encounters.

The results reveal a high potential for identifying unnoticed social interactions and encounters that may be important for mental health, especially among university students. While the findings hold promise for enhancing social interactions using technology, ethical considerations of privacy and consent in using geospatial data are also acknowledged.

This research establishes a new perspective on the convergence of technology, social interaction, and mental health, setting the stage for further investigations into this promising interdisciplinary field.