A proposed method for predicting US presidential election by analyzing sentiment in social media

Abstract

US Presidential election is an event anticipated by US citizens and people around the world. By utilizing the big data provided by social media, this research aims to make a prediction of the party or candidate that will win the US presidential election 2016. This paper proposes two stages in research methodology which is data collection and implementation. Data used in this research are collected from Twitter. The implementation stage consists of preprocessing, sentiment analysis, aggregation, and implementation of Electoral College system to predict the winning party or candidate. The implementation of Electoral College will be limited only by using winner take all basis for all states. The implementations are referring from previous works with some addition of methods. The proposed method still unable to use real time data due to random user location value gathered from Twitter REST API, and researchers will be working on it for future works.