Place visitation data improves health predictions by 7.5%

A recent study from the Penn State College of Earth and Mineral Sciences has shown that where people spend their time can be a stronger indicator of community health than where they live. Researchers led by geographers incorporated place‑visitation data—geographic points collected from millions of anonymous cellphone users with GPS‑enabled devices—into existing population‑health models. The addition of this mobility information raised the models’ predictive accuracy by an average of 7.5 percent.

The team’s approach involved mapping the daily movements of users across neighborhoods, workplaces, and recreational sites, then linking those patterns to health outcomes such as rates of chronic disease, vaccination coverage, and access to medical care. By capturing the dynamic nature of human activity, the enhanced models could better anticipate which communities might experience health disparities or outbreaks, offering a more nuanced view than static residential data alone.

These findings suggest that public‑health agencies could improve disease surveillance and resource allocation by integrating anonymized mobility data into their forecasting tools. As the use of location data becomes more widespread, such models may help target interventions more effectively and ultimately support healthier communities.