Visualization of Twitter Sentiments on Kashmir Territorial Conflict

Abstract

This paper aims to showcase the sentiments of Twitter users on the Kashmir conflict, which is an unresolved territorial conflict between two countries, India and Pakistan. Sixty thousand (60k) tweets have been collected using popular keywords posted by users living in the two involved countries and conflicting state. We proposed a twofold adaptive approach to exploit and compare different sentiment analysis techniques, namely lexicons (VADER and NRC) and machine learning. The analysis brings out that positive sentiments are higher in both the involved countries. However, there is a belief that people of both countries are negative and vicious about this act. The observations may help to break this myth and to foster positive engagement between the two nations, which are at the edge of a full-fledged war. The study provides a list of top Twitter influencer handles in each country and the disputed state. Thus, the purpose of this research is to scientifically analyze the situation to arrive at a fair decision and thereby removing unjust opinions that cause a lot of disturbance. This research can be used as fundamental to visualize the extensive insights of a critical event and help in identifying the groups which try to incite users on Twitter.