Analyzing the different sentiment of both the Covid-19 and H1N1 pandemics.
Both the Covid-19 (Dec 2019-) pandemic and the H1N1-09 (Jun 2009 – Aug 2010) pandemic have huge impacts on society. These events are covered by news outlets in real time and relay data and opinions. Sentiment analysis of such global health crises allows us to better understand public reception by analyzing a large number of news articles.
This framework can be used to perform sentiment analysis of any topic of interest. For our research, the range of a tone score was [0,1] and the score is only reported if it is more than .5.
As seen in Figure 1, the leading tones for Covid-19 are as follows: Sad, Analytical, and Tentative.
Both Sad and Analytical peaked around April of 2020, declining then remained stagnant for the following summer.
As seen in Figure 1 — Fear, Analytical, Sad, and Tentative peaked before the first wave around the summer of 2009.
Analytical, Fear, Sad, and Tentative were the prevalent tones.
It is important to note that the coverage for H1N1-09 is significantly less than that of the Covid-19 Pandemic.
Anger, Sad, Joy, and Fear weekly tone averages remained stable.
Tentative and Analytical possess the highest average at .67.
Higher fluctuations in tone scores, we suspect this is due to as previously mentioned problem of having much less coverage than Covid-19.
Sad was the most prevalent tone. There were no articles displaying the sentiment of Confidence. The Anger tone was seen ~3%. False Joyfulness was due to the double meaning of the keyword “positive” in IBM Tone Analyzer*
Sad, Analytical, and Tentative are the most commonly observed tones. The overarching tone seen in the articles though was Sad, which is most likely in reaction to the lives lost during these two time periods. Public sentiments peaked during the first wave of national lock downs in April of 2020 for COVID-19 but before the first wave for H1N1-09.
“IBM Watson Natural Language Understanding.” IBM, https://www.ibm.com/cloud/watson-tone-analyzer.