Whitespace teamed with epidemiological researchers in study find that Contact Metrics – which measure close interpersonal contact, a key driver of disease spread – better explain COVID-19 transmission at a municipal level than mobility data.
Whitespace Ltd., a spatial data analytics firm, and Dr. Forrest Crawford of Yale University have published a study in the Journal of Science Advances showing how mobile device geolocation data can be used to accurately predict COVID-19 cases in individual municipalities. Using geolocation data from mobile devices in Connecticut, the research team computed the frequency of close contact between individuals, aggregated the contact counts by week and location, and estimated the total intensity of contacts experienced by residents in each town.
The study accurately predicted when and where new outbreaks of COVID-19 were likely to occur. Close contact between individuals is a necessary condition for transmission of COVID-19. The team used data from mobile devices to compute the frequency of close contacts — within two meters (approximately six feet) — between individuals. The information was collected from February 2020 through October 2020. The study’s analysis showed that COVID-19 cases in individual municipalities could be predicted accurately using these contact metrics from 3 to 7 weeks prior, along with case counts.
The results indicate that allocation of testing resources in areas with both high prior contact and case counts may help detect or prevent emerging local outbreaks.
“The implications are significant for both state-level responses and resource allocation,” states Dr. Crawford. “Because we were able to show how data analysis is effective at the town level, can enhance individual municipalities’ capabilities to prepare for and deal with COVID-19 outbreaks.”
The study team gathered mobile device geolocation data as a proxy measure for physical distancing and movement patterns. This data is collected passively and was both anonymized and aggregated. Privacy and ethics in big data is a growing concern. The use of anonymized and aggregated data means that individuals are protected as no personally identifiable information is collected.
“Designing the study with privacy concerns in mind, we produced metrics that can’t be traced back to a specific individual. We’ve established that this approach is both ethical and effective,” added Jacqueline Barbieri, CEO of Whitespace Ltd. “That’s critical for putting the public’s mind at ease, while producing actionable results.”
For this study, Whitespace Ltd. applied a design approach and systems perspective to an inherently human problem, a technique that the company also has used to solve national security-related problems. The company is known for creating tools and methods to optimize sense making through Spatial Data Analytics.
The entire study is available in the Journal of Science Advances. More information can be made available by contacting the authors through Whitespace Ltd.
To read more about Whitespace’s journey with Contact Metrics, check out our previous blog posts.