Coronavirus has not only taken lives but it has taken away livelihoods as well. Loss of jobs and loss in businesses are other important casualties caused by a pandemic. Social media can be leveraged to chart the recovery of businesses and economic impact in nations ravaged by the COVID-19 pandemic. Researchers at the University of Bristol elaborate on a ‘real time’ method that has been accurately put to trial across three natural disasters worldwide. This method can be utilized in predicting the economic impact of the ongoing coronavirus crisis in various countries.
Interviews, surveys and many other traditional economic recovery estimates are generally time-consuming and costly. In addition, such methods do not scale-up well as well. Researchers from the Departments of Civil Engineering and Engineering Maths, University of Bristol were able to give accurate estimates using this method. They could predict the recovery of small businesses and downtime in countries damaged by three different natural calamities. They made use of aggregated data of social media to come with their predictions.
Analysis of Combined Public Posting Activity by Businesses to Help in Predictions
This ‘real time’ method relies on the assumption that businesses are inclined to publish more posts on social media when they open and less than when they are shut. Hence, analysis of the combined posting activity is done for a set of businesses over a period. This analysis helps in deducing accurately when the businesses are closed or open.
The researchers made use of the data available from the public posts of local businesses on Facebook. Those data were gathered after, during, and before the three natural disasters. Earthquake in Nepal in 2015, Hurricane Maria in Puerto Rico in 2017, and Chiapas earthquake in Mexico in 2017 were considered for the analysis. The findings of the analysis were validated with official reports, field surveys, analysis of Facebook post text and Facebook surveys along with other studies.
The framework of this new method works in ‘real time’ sans the need for text analysis and can be used on any type of natural disasters, including COVID-19. It can help local governments in optimum utilization and distribution of resources.
The findings of this new research have been published in Nature Communications, a scientific journal.