When the COVID-19 pandemic took hold, we wanted to get a better feel for the public discourse surrounding this topic. This seemed like a prime opportunity to look at what people were saying on social media platforms and how they felt about all the aspects of life that have been drastically changed by the pandemic.
The worldwide conversation surrounding COVID-19 offered a chance to apply Natural Language Processing (NLP) tools to analyze and find trends in what people were saying on social media. Using the Twitter API, we were able to scrape the text of millions of English-language tweets that mentioned COVID or Coronavirus. We then extracted key phrases from this massive body of text to find the topics being discussed most frequently. We also performed Sentiment Analysis to get a better idea of people’s feelings as they tweeted about these topics.
With this analysis completed, we then created a page to explore these topics. This includes visualizations to represent the conversation surrounding each topic and how the magnitude and sentiment related to a given topic and how it has shifted over time. We also included some sample tweets and related headlines for each topic, to provide additional context.
New tweets are scraped daily and added to our dataset so the list of extracted topics will change to reflect shifts in the conversation surrounding COVID. Explore it yourself at nlp.pandata.co!
Bob Wood is a Data Scientist at Pandata.