LAUSR

Abstract Gathering social media content and analysing the heavy and unstructured text coming from posts, comments and reactions can come as a powerful tool in understanding how people react to the information they receive. In this article we present the results from a social media analysis of 10771 headlines, with their subsequent text bodies and comments posted in a subreddit destined for Romanians during the state of emergency declared in Romania, from March 16 to May 15, 2020. Our objective was to model the main topics debated by this targeted population of people that tend to use Reddit to discuss current issues and to identify the sentiment polarity towards these topics. As expected, Romanians are mostly concerned with their social condition in the context of the pandemic caused by CoVID-19, as our research has revealed a word frequency for the term “Coronavirus” prominently higher than any other preferred term. However, the analysis brings up a surprising turnaround as the overall sentiment of the text posted in this dataset is predominantly neutral with a higher frequency of positive posts compared to the negative ones. This was unforeseen by our initial expectations: a natural tendency to more negative posts than positive considering the context of the chosen study period. Moreover, when compared to the time series of the CoVID-19 infections and caused deaths in Romania, spikes of extremely high or low mean sentiment scores per day can be correlated to the fluctuations of the declared cases. Not only does this bring us closer to understanding the social impact of CoVID-19 in the current context, but the outcome of this analysis can be easily extrapolated for further investigations upon other social networking tools or for more in-depth analysis on our studied corpus.