A social media platform using blockchain technology: Topic analysis and sentiment analysis of steemit user-generated content

Abstract

The "blockchain + media"model is showing great vitality. In this paper, we study the detection of the most common topics in Steemit. Compared to the currently popular centralized online social media, the blockchain-based decentralized online social media platform presents a different business model, and thus different or platform-specific topics emerge. We extract knowledge from content posted on Steemit, using NLP term frequency and topic modeling techniques combined with sentiment analysis theory, with the aim of studying the topic preferences and sentiment preferences of user-generated content. Our research will provide a more comprehensive understanding of identified themes and detect strategies adopted by users to add value to user-created content; provide guidance for business model improvements; and enrich existing interdisciplinary research.