Social Media Analytics during Pandemic for Covid19 using Topic Modeling

Abstract

The entire world is facing the Covid19 pandemic. This pandemic has various consequences on the political, cultural, economical and social life of the community. Lockdown has affected the psychological impact on society. This is reflected in various social media sites. In such a phase social media analytics for twitter data can be useful for understanding public opinion. In this paper, we have applied the Latent Dirichlet Allocation Algorithm as a topic modeling algorithm. Topic modeling finds the main theme that pervades the large data set. Twitter media is considered as the most popular microblogging platform, hence data during this pandemic is extracted from twitter. Natural language processing Techniques applied as preprocessing and then topic modeling applied which has given satisfactory results in terms of perplexity as a performance measure. Topic extracted gives an idea of the impact of Covid19 on society through their opinion on twitter. This can be helpful for making future policies by policymakers.