Please first read the following paper on the Covid-19 Twitter data. We plotted the spatial distribution of tweets mentioning Covid-19-related terms posted during February 19–22, 2020. Tweets with spatial coordinates represent a tiny fraction of all tweets (~0.5%). Still, there are enough of them to see national — and worldwide — trends.
First, at the national level, it is striking how much more early discussions there were in California about the Coronavirus than in the rest of the country. Interestingly enough, people in Washington state did not talk quite as much about it. Granted, its population is 5–6 times smaller than California, but given that Washington state was experiencing the effects of Covid-19 earlier and more intensely than any other state, due to a major outbreak near Seattle, it is surprising that there are not more tweets about it.
Another relative surprise is the limited amount of discussion observed in New York state during the time period under scrutiny. NY is the fourth most populated state in the US (19M), after California (39M), Texas (28M) and Florida (21M). The top three states are evidently the ones where most of the Covid-19 discussion occurred. However, NY doesn’t exhibit the same level of Twitter engagement. As of today (March 18, 2020), according to the CDC, NY has already reported over 2600 cases of Covid-19 infections, leading by a large margin, this unfortunate ranking.
With respect to the rest of the world, the US has many more discussions on Twitter about the Coronavirus. This is not surprising. What is interesting is Italy. The outbreak officially began in Italy on Feb 20, 2020. The very next day, the number of Twitter posts about the disease observed in Italy began to spike.
We will continue to probe the data to measure how people are responding to this unprecedented event.
Ashwin Rao, Emily Chen, Emilio Ferrara, Kristina Lerman