Title: Sentiment analysis of political discussion on Twitter in Nigeria's 2019 presidential election

Authors: Aderonke A. Oni; Samuel Oni; Blessing E. Udemezue

Addresses: Department of Computer and Information Science, Covenant University, Ota, 110001, Nigeria ' Department of Political Science and International Relations, Covenant University, Ota, 110001, Nigeria ' Department of Computer and Information Science, Covenant University, Ota, 110001, Nigeria

Abstract: Social media have helped in political communication by reinforcing social change and commitments from governments. This study analyses political discussion on Twitter during the 2019 Nigeria general election. The aim is to identify popular issues discussed and opinion of citizens on popular presidential candidates' social media network and correlate the online discussion with real world events during the 2019 election. Five most popular presidential candidates were selected for the investigation. Latent Dirichlet allocation (LDA) model was used to identify frequent terms and topics discussed. Sentiment of the tweets was analysed using National Research Council Canada (NRC) Emotion Lexicon. Findings revealed that Twitter is mainly used for political discussion that cantered on personal representation, political party's promotion and elections matters. The behaviour of users shows that Twitter platform is not being used to share original political ideologies, instead the platform was used for re-tweeting of tweets and political campaigns for candidates.

Keywords: social media; election; sentiment analysis; LDA; Latent Dirichlet allocation; emotion lexicon; Nigeria.

DOI: 10.1504/IJEG.2021.121241

International Journal of Electronic Governance, 2021 Vol.13 No.4, pp.462 - 485

Received: 27 Feb 2020
Accepted: 22 Mar 2021
Published online: 01 Mar 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article