Digital epidemiology: How to chart Covid-19 using search engine queries

According to Boris Ovchinnikov (a co-founder of e-commerce research agency Data Insight), the frequency of Internet searches for the word obonyanie (“sense of smell” in Russian) may serve as a metric for gauging the spread of the novel coronavirus across Russia. 

Since temporary loss of smell is a common symptom of virus, tracking the public’s interest in this term makes sense in the absence sufficient testing campaigns and fully reliable official statistics, the researcher explains in an op-ed in Vtimes.

“In mid-August (the two weeks from 10 to 23 August), the number of result impressions following queries for that term dropped to a minimum, 75,000 per week — which was 1.8 times less than at the May peak, but more than five times higher than average before the epidemic,” Ovchinnikov notes.

“Starting from the end of August, growth resumed and got faster week after week, reaching +40% week-on-week in late September and early October. “

“The next two weeks, growth slowed down, but still amounted to about 30% per week. The two last weeks of October, with 7% and 9% growth rates, suggested that the stage of exponential growth was over. However, so far we are not even talking about stabilization, but only a slowdown in growth,” Ovchinnikov concludes

Broken down by region, search results for obonyanie are likely show how Covid-19 moved and surged across Russia. For his research, Ovchinnikov relied on Yandex data, not Google data, because the latter offers only relative values — making it impossible to cross-reference different time periods and localities.

Topics: Data & Reports, Digital services & Apps, Search engines & SEO
Scroll to Top

This site is under maintenance. Sorry for the inconvenience.

This site is under maintenance. Sorry for the inconvenience.