Social
Measuring Human Mobility in Indoor Spaces: The MAPPING@Brown Study Jason R. Gantenberg* Jason Gantenberg Thomas A. Trikalinos Samuel F. Rosenblatt Guixing Wei Julia Netter Bradford Roarr Marta G. Wilson-Barthes Bryn C. Loftness Peyton B. Luiz Mark N. Lurie
Shortcomings in our understanding of disease transmission can hinder control of a circulating pathogen. Policymakers need detailed data on human mobility and interaction to inform decision-making and intervention design. During the SARS-CoV-2 pandemic, efforts to collect mobility data focused largely on the regional or national scale, but mounting effective responses to future pandemics will require supplementing these efforts with granular information on social mixing in indoor environments. We conducted the MAPPING@Brown study to explore the feasibility of collecting the relevant data. Specifically, we investigated the viability of using smartphone technology to infer individuals’ movements and contact patterns indoors. In Fall 2023, we deployed a network of 2,121 Bluetooth transmitters across 8 floors of the Brown University School of Public Health (SPH) building. Transmitters resided at fixed spatial coordinates such that the approximate maximum distance between a given smartphone and a transmitter would be 2 meters. In November 2023 we enrolled a cohort of 197 SPH students, faculty, and staff, whom we asked to complete a demographic survey and to download a custom smartphone application. This app recorded which transmitters an individual phone observed, a timestamp, and the received signal strength indicators (RSSI). Over the 2-week data collection period, we collected 138,073,742 Bluetooth signal readings from 128 unique smartphones. We will infer participants’ locations over time by estimating the distance between each smartphone and observed transmitter at a given time, based on a model for RSSI decay as a function of distance. Estimated distances will be used to derive devices’ spatial coordinates via multilateration and other methods. The resulting map of individual trajectories—complemented by auxiliary environmental data on CO2, temperature, and relative humidity—will inform high-resolution, spatially explicit agent-based models of indoor disease transmission.
