Once upon a time, say about 2012, Big Data and Massive Crunching was going to show us The Way. Better health, diagnosis, prevention, behavior, and a whole lotta other things. Doctors, nurses, engineers, and marketers feared that their jobs would be taken over by the handsome specimen to the left.
So at the start of the COVID pandemic, the hope was that Big Data was going to map the outbreaks and contact trace so that people could go into self-lockdown after a ride on a bus or subway, inform distancing measures, and identify hot spots for public health organizations, no matter how remote. Academic researchers and nonprofit partners mobilized into the non-profit Covid-19 Mobility Data Network that started by analyzing smartphone location data shared by tech companies. The intent was that public health officials could analyze it for insights based on hard data rather than 6′ guesstimates. It would then be expanded with additional data from Big Tech and grow, grow, grow.
Where it ran a cropper was the ad tech companies’ incompatibilities in data gathering, reluctance to share proprietary granular information, and privacy–an international battleground. Facebook turned out to be clueless in mapping mobility as a proxy or input to calculate contact rates, since it released only percent changes in movement or staying at home. The professors also didn’t figure on proprietary non-compatible systems and peculiarities stemming from business needs. Facebook, for instance, released data that mapped only eight-hour chunks in UTC which didn’t, of course, take into account normal bedtimes. Google would state that trends in staying home were up, versus Facebook data that indicated downward trends. Contact tracing, as Readers know, turned out to be a gigantic flop.
While the Covid-19 Mobility Data Network has evolved into a broader project called Crisis Ready, with the goal of creating data-sharing agreements that activate during a public health crisis, closing the gaps in data for epidemiological research remains elusive in areas such as urban versus rural and with specific demographics. STAT, PLOS Digital Health