Smartphone Location Data Can Leave Out Those Most Hit by Covid-19

Mobility knowledge from smartphones has helped authorities and healthcare officers throughout the nation make choices in battling the Covid pandemic, like whether or not to scale back capability at public locations or the place to position testing and vaccination facilities.

However the knowledge could also be flawed in lots of circumstances. A brand new research from researchers at Stanford College and Carnegie Mellon College means that smartphone location knowledge usually leaves out some individuals who could possibly be most in danger from Covid-19: those that are over 65 years of age and those that are nonwhite.

“Massive knowledge is just not all the time consultant knowledge,” says Daniel Ho, a legislation professor at Stanford and one of many authors of the research.

The researchers counsel that older individuals and folks of shade could also be much less more likely to carry smartphones or use apps that present their location. They in contrast smartphone mobility knowledge for election day in 2018 from geolocation agency SafeGraph Inc., which incorporates no data on private identities, with voter knowledge for that day from North Carolina, which incorporates details about voters’ race and age. The comparability confirmed that polling places the place extra aged and nonwhite individuals voted had been considerably extra more likely to be underrepresented within the mobility knowledge than places the place extra youthful and predominantly white individuals voted.

That discrepancy “has fairly important implications within the methods during which this knowledge and artificial-intelligence techniques educated on prime of it are used to tell pandemic response,” says Dr. Ho. “You might, as an illustration, miss sure nursing houses, to the extent that our research reveals that older populations are far much less more likely to be represented on this knowledge.”