On March 27, a man in Tucson, Arizona, posted on Twitter, "Guys. My wife is showing some symptons consistent with covid -19 [sic]. Seeing about testing."
The post is a vulnerable slice of life during the time of coronavirus, but data researchers say it may be something more. Mapping social media posts like this as data points could help municipalities to prepare for outbreaks.
Data researchers who have tracked the virus from Wuhan to Louisiana have identified 14 states they say might see their first spike in the next two weeks, based on a study of social media posts about personal experiences with the virus.
Transform talent with learning that worksCapability development is critical for businesses who want to push the envelope of innovation.Discover how business leaders are strategizing around building talent capabilities and empowering employee transformation.Know More Those states are: Alabama, Arizona, Indiana, Kentucky, Maryland, Minnesota, Missouri, Nevada, North Carolina, Ohio, South Carolina, Tennessee, Utah, and Virginia.
Dataminr, a New York firm that provides global real-time information to the United Nations, and multi-national corporations like 3M and McDonald's, has tuned into social media to plot firsthand public social media posts on COVID-19. The firm says the clusters of social media posts can be a leading indicator of virus outbreak hotspots from 7-15 days prior to exponential growth in COVID-19 official case count.
"These kinds of posts appear to be a leading indicator of where case count may be spiking in the future," says Dataminr CEO Ted Bailey.
In the United States, Dataminr found social media spikes from a week to two weeks before previous outbreaks in New York, California, Florida, Louisiana, Washington, New Jersey, Massachusetts, Pennsylvania, Texas, and Illinois. All of those states except Texas now have more than 4,000 cases and widespread community transmission, according to the US map used by the Centers for Disease Control and Prevention. The CDC did not immediately reply to a request for comment.
Digital media analyst Brian Solis praised the research. "Dataminr's approach to anticipating outbreaks is notable and welcomed," he said, and likened it to data from digital thermometer maker Kinsa, which is also being harnessed to forecast outbreaks. "Data is the answer to not only preparing in advance for medical needs but also in containing the virus."
The same trend also occurred in other initial COVID-19 outbreak countries, including South Korea, Italy, Spain, Iran, and France.
The UN, World Health Organization New York City Department of Public Health, and other public sector agencies use a Dataminr product called First Alert for breaking news alerts on the virus, the firm says. Companies are using the company's data to gauge potential impact on their workforce, supply chain, expenses, and decisions on travel and office closures.
The virus has now killed around 43,000 people, and infected around 875,000 people. Nearly 190,000 of those cases are in the US, which has the world's largest outbreak, representing more than one fifth of all cases worldwide. See all Business Insider's updates on the virus here.
A pattern that might be about to change
As of Monday there was an exponential growth in firsthand social media posts in the 14 US states named above that do not yet have a spike in COVID-19 cases. The CDC US map shows that none of those states have more than 2,000 confirmed cases. Ten of the 14 do not yet have widespread community transmission, marking a sharp contrast to the states that have seen the social media spike. (Arizona, Maryland, Utah, and Virginia do have widespread community transmission.)
If the Dataminr data pattern holds, that might be about to change.
The posts being measured include people indicating they tested positive for the virus, people indicating they are experiencing symptoms, people indicating they have been exposed but not tested, and first-hand accounts of confirmed cases from relatives, friends, and colleagues.
Tracking that messaging on Facebook, Twitter, Instagram, and other social media can show "clusters" of posts that indicate where a spike is coming - and give medical officials a chance to prepare supplies and other resources.
Predictions about COVID-19 are nothing new. Many medical experts have predicted different paths for the virus for months. Since February, researchers from the University of Massachusetts Amherst have conducted weekly surveys of two-dozen infectious-disease experts on the trajectory of the coronavirus outbreak in the US. The researchers' models offered a wide variety of outcomes, from a million US deaths this year to a few thousands.
In comparison to that research, social media's contribution to virus information has often been more suspect. The Pew Research Center found last week that people who get most of their news from social media follow virus news less and encounter misinformation more than those who get their news from other channels.
But what Dataminr shows is not people's opinions on social media. Its natural language processing computer programs sift through posts to find firsthand information of what people are going through in their own lives. Those posts pulled together form the clusters the company plots to show spikes in different areas.
There is precedent for tracking similar diseases on social media, as well. A 2017 study found a high correlation between mentions of the flu on Twitter and illness in the United States.
Bailey says the current isolation of many people around the world may place a greater emphasis on social media. "It is a time when people are recording what they're experiencing more, when people are communicating via social media more, and when people are contributing their experience to an unfolding global phenomenon more."
Social media has certainly had its struggles with misinformation, Bailey says. "But the ground truth is equally as powerful if you can use it to help with response."
Featured Health Articles:
- Telehealth Industry Explained
- Value-Based Care Explained
- Senior Care & Assisted Living Market
- Smart Medical Devices & Wearable Tech
- AI in Healthcare
- Remote Patient Monitoring Explained- AI in Medical Diagnosis Systems
Next Story
Global stocks rally even as Sensex, Nifty fall sharply on Friday
In second consecutive week of decline, forex kitty drops $2.28 bn to $640.33 bn
SBI Life Q4 profit rises 4% to ₹811 crore
IMD predicts severe heatwave conditions over East, South Peninsular India for next five days
COVID lockdown-related school disruptions will continue to worsen students’ exam results into the 2030s: study
