Discovering collective narratives shifts in online discussions

TL;DR

This paper introduces a systematic framework for discovering and analyzing the evolution of online narratives on social media, combining change point detection, semantic role labeling, and network aggregation, validated on COVID-19 and French Election data.

Contribution

It presents a novel computational method for automatically extracting and tracking narrative shifts in large-scale social media data, addressing a key challenge in understanding online discourse dynamics.

Findings

Successfully identified major narrative shifts aligned with key events.

Demonstrated effectiveness on synthetic and real Twitter datasets.

Revealed insights into how narratives emerge and evolve over time.

Abstract

Narrative is a foundation of human cognition and decision making. Because narratives play a crucial role in societal discourses and spread of misinformation and because of the pervasive use of social media, the narrative dynamics on social media can have profound societal impact. Yet, systematic and computational understanding of online narratives faces critical challenge of the scale and dynamics; how can we reliably and automatically extract narratives from massive amount of texts? How do narratives emerge, spread, and die? Here, we propose a systematic narrative discovery framework that fill this gap by combining change point detection, semantic role labeling (SRL), and automatic aggregation of narrative fragments into narrative networks. We evaluate our model with synthetic and empirical data two-Twitter corpora about COVID-19 and 2017 French Election. Results demonstrate that our approach can recover major narrative shifts that correspond to the major events.