The new eye of smart city: Novel citizen Sentiment Analysis in Twitter

Abstract

Governments across the world are trying to move closer to their citizens for better smart city monitoring and governance. Twitter Sentiment Analysis is opening new opportunities to achieve it. In this paper, a methodological framework to collect, pre-process, analyse and map citizen sentiment from Twitter in helping the Governments monitor their citizens' moods is proposed based on the prior works. Multinomial Naïve Bayes classifier is used to build a sentiment classifier, which employs a variety of features including a specific microblogging feature - emoji. Our proposed sentiment model outperforms the top system in the task of Sentiment Analysis in Twitter in SemEval-2013 in terms of averaged F scores. The novel feature emoji has proved to be useful for Sentiment Analysis in Twitter data in this work. We also apply our model to real-world tweets and present how Government agencies can track the fluctuation of citizens' moods using mapping techniques.