The escalation in social media usage amid the pandemic has gathered extensive data for Natural Language Processing tasks, many of which solve some of the contemporary issues of the world. An abundance of these NLP tasks focuses on widely spoken languages like English. Moreover, researchers haven’t exploited the true potential of NLP and the available data to extract a solution for prominent issues like the COVID-19 epidemic. This project, presents a COVID tweets analysis platform that allows health decision-makers to view a real-time analysis of narratives of tweets, particularly in Nepali and Devanagari scripts. The method used here enables stakeholders to look into the public conversation and trends surrounding the COVID-19 epidemic. It categorizes the tweets across nine curated (mapped from WHO) topics using a MuRIL model to summarize our insights. Experimental results reveal the essence of the model as it can classify the tweets with appropriate scores for the evaluation metrics. The analysis of the results shows that the model can portray the growth in the discussions relating to different phases of the epidemic.
Publiction URL:
https://www.researchgate.net/publication/361363246_Epidemiological_Surveillance_System_using_NLP
Project Github:
https://github.com/naamiinepal/covid-tweet-classification