Topic Modelling Twitterati Sentiments using Latent Dirichlet Allocation during Demonetization

Twitter has surfaced as one of the major social media platforms for sharing political views on pressing issues for the common man. In this paper, we attempt to apply a topic modelling technique, namely Latent Dirichlet Allocation (LDA) on tweets to analyse and come up with pertinent topics with the most relevant words that describe the topics most aptly. The tweets contain the demonetization hashtag to help us understand sentiments of people about demonetization. This technique can be used by analysts across any industry to understand the pertinent topics revolving around any social issue to take further actions in their organizations. Leveraging this sort of statistical topic modelling can be quite useful to researchers to correctly identify primary components of huge textual corpora for any kind of further analysis. The model comes up with very meaningful categorization of topics. Further, we measure the inter-topic distances via multidimensional scaling and review words and topics through metrics such as saliency and relevance.