Technologies for social media narrative analysis: review

Annotation: Social media play a pivotal role in modern communication. The primary element of textual content in social media is narratives, which convey sequences of events and reflect the author’s perspective. As the volume of user-generated data grows, the task of automated narrative extraction from text is becoming increasingly critical. The goal of this article is to conduct a comparative analysis of existing methods, approaches, and tools for automated narrative extraction, followed by a formalization of their structure. Such formalization is essential for algorithmizing the narrative extraction process using NLP tools (Natural Language Processing). The study explores key approaches to narrative representation, including methods based on bag-of-words, semantic annotations, ontologies, and vector representations (word embeddings). Their limitations and domains of application are analyzed. The special emphasis is placed on adapting these methods to analyze short, fragmented texts typical to social media (e.g., posts, tweets). The findings of this research useful for the development of narrative analysis algorithms, offering new opportunities for applications in marketing research, public opinion analysis, and psycholinguistics.

Keywords: review, social media analysis, natural language processing, computational linguistics, computational narratology, narrative modeling, narrative