/COVID-TwitterDiscourse

Final Project- DS-GA-1003

Primary LanguageJupyter Notebook

COVID-TwitterDiscourse

Final Project- DS-GA-1003

We are interested in the analysis of political discourse onTwitter in the US with respect to the ongoingCOVID-19pandemic. The motivation behind this analysis is toidentify political narratives that harnessed more sup-port among the users of the micro-blogging site andthat were pushed by different news agencies owing totheir implicit political bias. To build the training cor-pus, we will utilize tweets pertaining toCOVID-19madeby the US Governors hailing from Democratic or Re-publican parties, allowing us an efficient way to identifyand label tweets representing discourse pushed from ei-ther side of the political spectrum to deal with the samecrisis. Using the trained model, we intend to analyze theTwitter accounts of all major US news outlets to iden-tify the support for either of these political discoursesand whether their stance evolved over time. For possiblefuture work, another application of the model could beused to track real-time twitter sentiments in responseto real-world events such as, how do the postings onthe platform evolve following events, such as the WhiteHouse briefings.