Polarity : responses to coronavirus where the most polar at the beginning of the out break and shows an overall trend of becoming neutral as time goes by. Subjectivity : Overall same high subjectivity score of 0.35~ 0.4 is maintained.
Tweets on the economy increased on March 16th when the stock market met a plunge. Overall countries had similar amounts of positive and negative responses to the economy. We've seen increases increase in spikes of economic tweets when stock market changes and border restrictions to countries were announced. Overall there were more negative responses to the tweets.
Swiss and Spain have the most dramatic differences between the amount of positive and negative terms used. For these countries approximately 70% to 80% percent of the tweets about the economy are negative For other countries, there isn't a dramatic difference between the amount positive and negative terms used. However, we could find a overall trend of negative terms more frequently used.
To test the accuracy of our models, we used the one-way ANOVA test which test for analyzing the variance of the groups to see if there is a
difference between them. Based of the similarity matrix, we created 8 clusters and made polarity/subjectivity calculations for each of the groups. Based off
that, we are trying to see if the polarity/subjectivity calcuations for each of the groups have a significant difference.
Ho: The polarity/subjectivity scores for the eight clusters do not differ
Ha: At least one of the polarity/subjectivity scores for each of the eight clustes differ.
As a result, we got a p-value of 0. Therefore, we reject the null hypothesis and conclude that the polarity/subjectivity scores for at least one the eight clusters differ
