Segregation and User Interactive Visualization of Covid-19 Tweets Using Text Mining Techniques

One of the worst calamities the world is facing since early 2020 is corona virus or Covid-19 disease which has turned into a pandemic claiming millions of lives across the globe. Twitter sources huge number of tweets related to this disease from users globally. This research focuses on mining Covid-19 tweets using machine learning techniques. The tweets are first pre-processed and converted to a form suitable for applying clustering algorithms. Principal Components Analysis is used to separate most significant components. Similar tweets are categorized using Hierarchical agglomerative clustering. The segregated tweets are visualized on novel and interactive cluster plots, members of which can be identified on user interface interactively by user for easy interpretation. The implementation is done using R programming. Clusters of similar tweets can be used to analyze the response of people to the pandemic across countries, compare and adopt best practices across countries to address the pandemic based on people views, combat spread of rumors and other such applications.