- A team of researchers analyzed Twitter mentions of pneumonia and dry cough, two common coronavirus symptoms, to determine the arrival of COVID-19 in Europe.
- The scientists found a statistically significant increase in mentions of COVID-19 symptoms in early 2020, weeks before the first cases were confirmed in several European countries.
- Using geolocation data, they also observed an increase in COVID-19 symptoms mentions in locations that would later become major coronavirus hotspots in Europe.
More than a year after the first cases appeared in Wuhan, we still don’t know how the COVID-19 pandemic started. The World Health Organization is in the midst of an official investigation into the matter, which started only this month — also more than a year later than we expected. The scientific consensus is that the illness started in an animal, then spread to an intermediary host or directly to a human. But China hasn’t helped move the needle in any direction, although it has investigated the matter. In fact, China has shown more interest in obfuscating the origin story, looking to avoid taking the blame for the pandemic. Beijing has offered several alternate COVID-19 origin stories, with the latest one having been divulged just days ago. China suggested the pandemic might have started in America, in response to Mike Pompeo’s recent claim that the virus was a lab accident.
We’ve learned in the past year that COVID-19 arrived in various countries much earlier than the first PCR test performed in that nation. Studies from France, Italy, and the US showed the virus was spreading in late 2019 in some regions, at a time when the virus was emerging in Wuhan. China used some of these studies to imply the virus evolved in a different region and then arrived in Wuhan, without providing any evidence to support such hypotheses. The other explanation might be that the virus got out of China much earlier than anticipated.
Now, a team of researchers is taking an interesting approach to study the origin of the pandemic.
They looked for mentions of pneumonia on Twitter, which often arises in severe COVID-19 cases. Another reason why the scientists used the coronavirus symptom was the fact that the flu season was significantly milder in 2020 compared to previous years, so the flu would be a less likely explanation for pneumonia reports on social media.
The authors looked at mentions for pneumonia in the seven most spoken languages in Europe (English, German, French, Italian, Spanish, Polish, and Dutch) between December 2014 and March 1st, 2020, making a few key corrections. Tweets containing the word pneumonia from news reports from December 31st, 2019, the day when the mysterious illness was announced, to January 21st, 2020, the day when the human-to-human spread was officially announced, were removed from the database.
The authors found an increase in tweets mentioning pneumonia in most European countries in the study as early as January 2020, which showed ongoing concern and public interest in pneumonia cases.
Italy was one of the first European countries to experience a massive first wave of COVID-19, with the first local case detected on February 20th. But the Twitter mentions of pneumonia spiked in the first weeks of 2020 compared to the same period a year earlier. This seems to indicate the existence of COVID-19 hotspots before the illness was officially diagnosed.
France saw a similar pattern, while Spain, Poland, and the UK got a delay of two weeks.
The researchers also used the geotagging data in tweets mentioning pneumonia to find potential hotspots. Unsurprisingly, the tweets came from the regions that would report major COVID-19 outbreaks later, including Lombardia in Italy, Madrid, Spain, and Île de France.
The authors then repeated the steps for a different significant COVID-19 symptom, dry cough. They observed the same pattern in tweets, an abnormal and statistically significant increase of mentions of the symptom in the weeks leading to February 2020. This indicates that the virus was spreading in Europe much earlier than anticipated.
Studying social media for the emergence of unusual symptoms might be a way to pinpoint the increasing threat of a new infectious disease like COVID-19. Twitter might not explain how the coronavirus infected Patient Zero, but it could determine when the illness arrived in various locations. The study only covers Europe, but similar research could be done for the US and other world regions. A similar analysis of Weibo data could indicate whether people in China complained about pneumonia and dry cough in the months or weeks leading to the main Wuhan epidemic.