Facebook is starting to share more location data with COVID-19 researchers and asking users to report symptoms themselves
Facebook is expanding a program that will give researchers access to movement pattern data in an effort to improve our understanding of the spread of COVID-19, the company said today. Data for good, which uses aggregated, anonymized data from Facebook apps to inform academic research, now gives access to three new maps to predict the spread of the disease and reveal whether residents of a particular region will stay at home.
The company will also ask Facebook users to take part in a Carnegie Mellon University survey asking people to report disease symptoms themselves. The responses, which will be anonymized, can help researchers understand new hotspots as they develop or see where the disease begins to withdraw. Carnegie Mellon does not share symptom information back to Facebook, the company said.
Last week, Google released public reports that use the company’s own location tracking services to reveal the extent to which people have changed their movement patterns in response to the global pandemic. Facebook had comparable information has already been made available to academic researchers.
The moves announced today are designed to improve forecasts and response efforts in the United States and beyond. In an interview, Facebook executives said the company can help repair illness while protecting the privacy of individual users.
“We think Facebook and the broader technology industry can and should continue to find innovative ways to help health experts and authorities respond to the crisis,” said Steve Satterfield, director of privacy and public policy at Facebook. “But we don’t think these efforts should compromise people’s privacy. We think we can help respond to public health while protecting people’s data.”
The tools released Monday include colocation maps, illustrating the extent to which people living in different areas mix; range of motion trends, which show the extent to which people stay at home or go out; and a “Social Connectivity Index,” which shows how likely two people are to become Facebook friends, a measure of the strength of social bonds in a particular place. Communities with stronger social bonds may recover faster than others, said Laura McGorman, Data for Good policy leader.
The disease prevention maps are based on data from Facebook that has been processed to hide individual identities, the company said.
“Measuring the impact of social distance policy is absolutely crucial at this stage, and such aggregated data provides insights that protect individual privacy but are useful for policy makers and researchers who build predictive models,” said Caroline Buckee, associate director of the Center for Communicable Disease Dynamics at Harvard’s TH Chan School of Public Health, in a statement.
Andrew Schroeder, who runs analysis programs at the humanitarian aid organization Direct Relief, said the new maps would help public health organizations and aid groups understand the effectiveness of those staying at home and help plan response efforts. The data could inform changes in public health officials’ reporting and any plans to end current locks, he said.
Schroeder said that mapping efforts by major technology companies helps researchers understand the effectiveness of social distance guidelines in real time, improving models that track the spread of COVID-19. “It was three weeks ago, ‘Is anyone staying home?'” He said. “Now is it true? How? How much? How much is enough? How does this affect the number of cases? That’s the agenda.”
Facebook also puts a prompt at the top of the news feed in the United States and invites users to report disease symptoms themselves to Carnegie Mellon’s research center in Delphi. Given the persistent widespread shortage of tests across the country, reports of symptoms may provide an imperfect but still valuable view of where the new coronavirus can spread before public health officials are aware of it. Facebook shares everything you report with a random identification number, along with a statistical weight value that corrects for bias in the sample. (If different communities respond in different numbers, Facebook says, the statistical weight will explain that.)
Researchers can apply to access Data for Good website. Facebook works with more than 150 universities and non-profit organizations to date, McGorman said.