An Application Of Sentiment Analysis With Transformer Models On Online News Articles Covering The Covid-19 Pandemic

This masters thesis by Prakul Asthana at the University of California, Los Angeles (UCLA) uses GDELT as a news index for sentiment analysis:

The Covid-19 pandemic has had a devastating impact on lives across the world, with tremendous human socio-economic costs, while exposing and exacerbating several fault lines in our society. It has also caused a rapid rise in misinformation and erosion of trust in established news outlets amid allegations of political bias and censorship. In this paper we use the processes of sentiment analysis to study the coverage of the Covid-19 pandemic in news outlets. By comparing the coverage from news sources with opposing political leanings, we quantitatively establish political bias. We also repeat this process on news articles mentioning specific topics like Masks, Social Distancing etc., to check for any bias present in the sentiment towards them. Lastly, we compare sentiment in Covid-19 news coverage in the United States, the United Kingdom and Australia to contrast the political bias in news articles on the pandemic in these three countries.

Read The Full Thesis.