The spread of coronavirus can be partially predicted via Facebook friendship patterns, a New York University study has concluded.
NYU researchers tracked the progression of the disease from two early hotspots in suburban New York state and Lodi province in Italy, finding the spread was strongest in areas that with significant links to the regions through the social network.
For instance, in Lodi the provinces with highest COVID-19 case densities and connectedness are in the surrounding Lombardy region, as well as the nearby Piemonte and Veneto regions. There are also relatively high levels of both connectedness to Lodi and coronavirus cases in Rimini, a popular tourist destination along the Adriatic sea. A number of provinces in southern Italy send workers and students to the industrial Lombardy region, and therefore have strong social ties to that region.
While some of these areas have seen a number of cases, they are not disproportionally larger, perhaps reflecting the efforts of Italian authorities to restrict the movement of individuals. This tends to suggest not all connections hold, although the team say the Social Connectedness Index appears to have predictive power above these other measures that might commonly be used to proxy for social interactions.
The Social Connectedness Index is a tool built in conjunction with Facebook, a dataset which allows researchers to measure the connectedness of two geographic areas as represented by Facebook friendships, without needing to access raw social graph information itself, which could infringe user privacy. The resource can be accessed by other academics and non-profits wishing to conduct similar research.
In suburban New York county Westchester, “coastal regions and urban centres appear to have both high levels of connectedness to Westchester and larger numbers of Covid-19 cases per resident”. Specific hotspots could also be found in popular holiday locations frequented by “well-heeled residents” of the county, such as coastal Florida near Miami and ski resorts in southern and central Colorado.
“This finding suggests to us the geographic structure of social network as measured by Facebook may indeed provide a useful proxy for the type of social interactions epidemiologists have long known to contribute to the spread of communicable diseases,” the researchers say.
Massive datasets offer much opportunity to epidemiologists at the beginning and end of disease outbreaks, when the number of new cases of is low but rising or high but declining, as they could reveal able which geographical areas are most at risk when a new cluster forms.
Facebook has made a number of these massive datasets available to researchers under its Data For Good programme, but has so far avoided handing specific user data to governments and public health authorities directly.
