a Group of scientists at Harvard University have developed a mathematical model that based on data from social networks and search engines are able to predict outbreaks of coronavirus infection COVID-19 approximately two weeks before they occur. This time is sufficient for the authorities to quickly adopt measures to curb the disease, reports The New York Times.
the Algorithm created by the team led by Mauricio Santillana and Nicole Kogan, a real-time analysis of Twitter, the search queries in Google, geolocation on smartphones, as well as data from other sources. On the basis of this information, builds a model that can apply early signals about the growth of infections COVID-19.
a Similar tool, Google Flu Trends – since 2008 developed by the company Google. Instead of relying on the official statistics about the number of cases, this web service predicts outbreaks of flu based on the growth in the popularity of certain search terms (e.g., “feel very tired”, “joint pain” or “dosage Tamiflu”).
As noted by the NYT, the new draft goes further, adding the Google searches of other types of data: with social media posts on the topic of coronavirus, a request for a medical platform for doctors UpToDate, anonymous data from smartphones and even indicators with smart thermometer Kinsa, which is loaded into the application. In the future, it is planned to connect to the online store Amazon and video service Netflix.
if a new system out soon. The first forecast algorithm is done: it is expected that the States of Nebraska and new Hampshire, the number of new cases COVID-19 in the coming weeks, will increase, unless more decisive measures.