The Canadian BlueDot software had detected the Coronavirus outbreak several days before the WHO.

This algorithm continuously scans the news in 65 languages, the various and varied reports on all diseases, infections and other small scratches. It is able to sort out when someone is talking about news and someone complaining about coughing or fever on social networks. He is especially able to diagnose on the Web if people are suffering from a common flu, a cold or if a new virus appears by analyzing the symptoms.
By adding up all this data, it can discern a signal that can be compared to a needle in a haystack.
By having access to data on the movement of infected people, he is able to map an infectious outbreak and, above all, to predict its geographic spread.
Of course, all these data are validated or invalidated by epidemiologists who analyze the results.
In the case of Wuhan, BlueDot predicted the evolution towards Bangkok, Seoul, Taipei and Tokyo in the days following its appearance.
The start-up had sent its findings free of charge to health authorities, airlines and public hospitals, but faced with the risk of being misled, these alerts did not receive the reception they deserved.

Online retail giant Alibaba, which set up its research institute a few years ago, has analyzed data from more than 5,000 confirmed Covid-19 cases and has perfected its artificial intelligence algorithm to detect cases of infection in less than 20 seconds, with an accuracy rate of up to 96 per cent.
Ping An, a health organization that has developed a similar algorithm, has a 90% efficiency rate and detects infection in 15 seconds.
Conventional techniques require the analysis of more than 300 images and take a minimum of 15 minutes.

Note the late entry of Google, Facebook, and other U.S. technology companies that are looking at the potential use of personal data to fight the epidemic. This would involve collecting location data from American smartphones to map the spread of the disease and predict future urgent medical needs.

M.D