Twitter data analytical methodology development for prediction of start-up firms’ social media marketing level

Abstract

Social media marketing is an essential and important tool for start-up firms, which can help start-up firms remedy the marketing limitations through ease and relatively low costs. Predicting start-up firms’ social media engagement level can allow them to gauge the effectiveness of their social media marketing efforts and can provide numerous benefits related to strategic marketing processes. This study focuses on developing a methodology involving data science processes and machine learning models to account for the ongoing advancement of business intelligence methodologies. This study gathered data of 8,434 start-up firms from Twitter, generated social media-based features, and created machine learning models to predict the social media engagement level of each firm. The results show that deep learning provides the best accuracy in predicting the engagement levels. The results also show that the number of tweets by the firms, the number of retweets received, and the number of likes received have the most significance in determining the effectiveness of social media marketing activities.