Twittering Public Sentiments: A Predictive Analysis of Pre-Poll Twitter Popularity of Prime Ministerial Candidates for the Indian Elections 2014 Suresh Kalyani*, Ramakrishnan Chitra Amrita Vishwa Vidya Peetham, Coimbatore, India *Correspondence to: Kalyani Suresh, Amrita School of Communication, Amrita Vishwa Vidyapeetham, Ettimadai, Coimbatore-641 112, Tamil Nadu, India. E-mail: s_kalyani@cb.amrita.edu
Online published on 12 September, 2016. Abstract Twitter is a useful tool for predicting election outcomes, effectively complementing traditional opinion polling. This study undertakes a volume, sentiment and engagement analysis for predicting the popularity of Prime Ministerial candidates on Twitter as a run-up to the Indian Elections 2014. The results from a survey of 2, 37, 639 pre-poll tweets finds tweet volume as a significant predictor of candidate vote share, and volume and sentiments as predictors for candidate engagement levels. Higher engagement rates evolve from the horizontality of conversations about the candidate, therefore indicating a high degree of interactivity, but do not translate into a higher vote share. Top Keywords Twitter analytics, Indian elections 2014, Modi, Kejriwal, Rahul Gandhi, sentiment analysis, twitter engagement rate. Top |