Decoding semantic news ecosystem: Macro centrality analysis in Twitter news entity graphs and global influences

Abstract This research holds significance in unraveling the intricate interplay between global events and the semantic space of the news ecosystem. This investigation is motivated by the need for a comprehensive understanding of the structural dynamics and information flow within the Twitter news ecosystem. This research uniquely contributes to the existing literature by providing a longitudinal analysis of the macro-scale features of Twitter news entity graphs, specifically exploring their nuanced relationship with global events and the evolution of the news ecosystem. Over a span of 12 years and across 55 prominent news channels, this study employs a robust methodology, utilizing natural language processing techniques to extract named entities. These entities form the basis for constructing daily graphs which represent the semantic space, facilitating the analysis of co-occurrence patterns and macro centrality measures. The computed macro centrality measures undergo trend analysis, revealing temporal patterns and evolution. Subsequently, extremum analysis is applied to discern the impact of global events, while Fourier and wavelet techniques in frequency content analysis enrich the understanding of macro feature characteristics. Situated within pertinent theoretical frameworks, these findings offer insights into the profound influence of global events on news semantic space and inform practical applications, including event detection, prediction, and combating misinformation. Such insights into news entity graph dynamics can further assist news organizations in strategic planning and adapting to the ever-evolving digital landscape.