Cross-language sentiment analysis of European Twitter messages duringthe COVID-19 pandemic

Social media data can be a very salient source of information during crises.\nUser-generated messages provide a window into people's minds during such times,\nallowing us insights about their moods and opinions. Due to the vast amounts of\nsuch messages, a large-scale analysis of population-wide developments becomes\npossible. In this paper, we analyze Twitter messages (tweets) collected during\nthe first months of the COVID-19 pandemic in Europe with regard to their\nsentiment. This is implemented with a neural network for sentiment analysis\nusing multilingual sentence embeddings. We separate the results by country of\norigin, and correlate their temporal development with events in those\ncountries. This allows us to study the effect of the situation on people's\nmoods. We see, for example, that lockdown announcements correlate with a\ndeterioration of mood in almost all surveyed countries, which recovers within a\nshort time span.\n