Chinese Economist Uses Baidu Search Data to Forecast Future COVID Cases
A popular economist on Chinese Q&A platform Zhihu released a report on December 11 regarding the initial round of expected mass COVID-19 infections since the government lifted tight restrictions in recent weeks. The forecast, published by a verified account called “chenqin,” is based on Baidu search data for “fever,” showing that many cities in China may reach their peak before the Spring Festival in late January.
The economist was previously a university teacher engaged in economic research, and now works as the chief economist at a big data company. He has been dealing with government data for many years now. He divides Google search index into pandemic period and non-pandemic period. Processing the average “fever index” during the non-pandemic period and the search index of fever during the pandemic period allows him to draw conclusions about the spread of the virus.
Specifically, if the proportion of people making searches about fever and the proportion of people catching fevers are relatively stable, then the cumulative area of excess fever search index is directly proportional to the proportion of the infected population to the total population.
The economist found that in Taiwan, Hong Kong and Japan, when infections reached their peak, the cumulative area of excess fever search index all reached 80. At the end of the first wave of outbreaks in two regions, the index reached 200, 160, and 250 respectively.
The economist then conducted a similar study with the Baidu search index, taking 100 as the cumulative area of excess fever search index when the infections reached their peak, and calculated the peak time and end time of the outbreak in every city.
He listed all the cities that can reach the infection peak before the Spring Festival and the proportion of people who have been infected before the peak in these cities. For example, in Beijing, as of December 13, the peak of the first wave of group infection has been completed by 47%, and the peak of the first wave of infections is on December 16, and the ending date is January 4 next year. In Guangzhou, the three indicators are 10%, December 24 and January 20 next year.
On December 10, the economist posted that there were at least 360 million people in China who are victims of group infection. From the experience of other countries, it is expected that all cities in China will enter the group infection in December, and the first wave of infection will peak in four to six weeks, then start to decline. Therefore, some cities can end the first wave of group infection earlier, such as Xingtai and Shijiazhuang, which have basically reached the peak of infections.
Beijing and other places are expected to peak from the end of December to the beginning of January, and after January, new cases will gradually decline. If a city has not seen mass infections by the end of December, it will probably hit a peak during the Spring Festival period.
SEE ALSO: Chinese COVID Pill Azvudine to Open for Online Sales
Baidu, the leading search engine in China, is also using public search keywords to monitor large-scale changes and living conditions in China. Baidu search data has revealed that searches for COVID prevention have surged 15-fold in the past week. In particular, items such as N95 masks, ibuprofen and Lianhua Qingwen, a traditional Chinese medicine formulation used to treat influenza, have received a lot of attention. Additionally, terms such as “job change,” “back to work,” and “returning to school” have become popular keywords. These findings suggest that people are increasingly concerned about both their health and their future plans.
Baidu also listed the cities that have gradually returned to normal life according to the travel data of Baidu Maps. The top five cities are Urumqi, Xi’an, Nanjing, Taiyuan and Zhengzhou.
