Google vs. COVID-19

In the middle of March 2020, the world felt like a hellscape under siege. After being declared a pandemic by the World Health Organisation (WHO), the COVID-19 outbreak had dumbfounded public health officials and policymakers alike by putting them on the spot. Besides being logistically ill-equipped, governments did not have the general wherewithal to formulate policy responses, and before we knew it, an air of paranoia had compelled numerous economies to enter a state of lockdown.

At the time, however, authorities worldwide were unable to acquire any semblance of consolidated information to gauge the effectiveness of the mobility restrictions. It was at this point that Google, the world’s most popular search engine, rose to the occasion. Ostensibly, it had dawned on Google quite early on that the aggregated, anonymized insights collected through applications like Google Maps had the potential to help policymakers assess mobility trends. In April 2020, Google made the call to publish the data it collects in the form of reports. Google’s Mobility Reports have been in the public domain ever since and chart movement trends both temporally and spatially. Google compares mobility patterns across six categories: retail and recreation facilities, grocery stores, pharmacies, parks, transit stations, workplaces, and residential areas. For the sake of simplicity, a seven-day moving average is used. The information collected before February 23, 2020, for each day of the week serves as the baseline for comparison for the corresponding day of the week.

Initially, Google’s COVID-19 Community Mobility Reports covered 131 countries. Recently, however, Google went the extra mile to make the reports all the more accessible. The reports are now updated thrice a week in 64 languages, with hyperlocal insights covering 13,000 regions across 135 countries. By using aggregated, anonymized data from android users who have enabled the Location History feature, Google’s COVID-19 Community Mobility Reports strive to paint a picture of what has changed as a result of mobility restrictions.

At their core, Google’s Mobility Reports seek to identify, quantify, and classify both short- and long-term implications of local lockdowns. While no source of publicly available data has the ability to get down to the nitty-gritty of certain temporal elements, Google lets users use algorithmic filters to help them make sense of the data. For instance, experts have the option of deploying the Kálmán filter to understand mobility trends across the globe. Invented by Rudolph E. Kálmán, the eponymous filter can utilize noisy measurements to estimate the state of unobservable variables, making them ideal for dynamic systems. What’s more, Google’s Mobility Reports provide individuals with footfall predictions. As a result, the reports empower businesses to set their opening hours more effectively.

Since local lockdowns had been nothing short of an exogenous shock to the world economy, Google’s COVID-19 Community Mobility Reports had already become a closely watched economic indicator by June 2020. One must note that this endeavor, howsoever well-meaning, was bound to raise grave concerns about the privacy of individuals at some point or other. It is Google’s willingness to embrace transparency that makes all the difference.

Much to the world’s relief, Google made it clear that it has been committed to treating personal data responsibly from the word go. The privacy of individuals remains protected by dint of the stringent protocols in place. Depending on the product in question, Google acquires certain data elements. Even though it is easier to link some of these data elements to specific individuals, the generalization techniques used by Google work a treat by concealing and anonymizing the identity of users. Additionally, the internet giant makes use of differentially-private algorithms to add mathematical noise to the data it generates. Because these algorithms throw random bits of noise into the mix, they render the data immune to adaptive attacks that rely on auxiliary information. In other words, no personally identifiable information is made available at any point.

As one continues to dig into Google’s massive trove of data, it becomes increasingly apparent that these reports can be used in conjunction with other epidemiological tools to track the spread of COVID-19. Google has managed to pick up some brownie points by announcing that it is looking to collaborate with epidemiologists to prepare a dataset to forecast the trajectory of the pandemic. Of late, Google’s Mobility Reports are being hailed as an indispensable epidemiological tool the world over, and for good reason.