Online social media platforms such as Twitter and Facebook offer a flexible and effective means of communication on a large scale leading to a vast amount of diverse datasets. This capability makes it possible to analyse large amounts of data virtually on any topic from online users. Organisations, media houses, politicians, artists, entertainers, among others, utilise social media to engage with a broader audience. Social media platforms play an instrumental role in election campaigns and other political debates, as reported in past studies. Despite the growing number of studies concerning the role of social media in elections, little attention has been paid to Africa's largest democracy (Nigeria); this motivates our study. For the study, we collected a large number of tweets based on the Twitter handles of the political parties and presidential candidates. We measured the online activities of political parties and candidates and analysed users' opinions to understand their sentiments. Moreover, we used differential sentiment analysis to study the relationship between the opinions of the users and the information produced by the 12 top Nigerian news media outlets. We apply a novel method of reverse engineering of Twitter accounts to understand the online campaign strategies of the parties. We found that while both the ruling party and minority parties concentrate on promoting their candidates, the top opposition party plays a two-way strategy of promoting their candidates while attacking the ruling party and its candidate. Finally, compared to the actual election results, our Twitter analysis correctly predicts the winner in 24 states out of the 37 states (including the Federal Capital Territory). Our findings provide insight into the 2019 Nigerian election from social media perspectives and encourage further research about the country using data from social media.
