Quantifying Information Distribution in Social Networks: The Structural Entropy Index of Community (SEIC) for Twitter...

Area of Science:

• Computational Social Science
• Network Science
• Information Theory

Background:

• Social network analysis (SNA) often uses synthetic data, limiting real-world applicability.
• Understanding online political discourse requires robust analytical tools.
• Existing methods may not fully capture the complex dynamics of digital social networks.

Purpose of the Study:

• To present an integrated approach combining graph theory, SNA, and information theory for analyzing real-world social networks.
• To introduce a novel metric, the Structural Entropy Index of a Community (SEIC), for quantifying community decentralization.
• To analyze a live Twitter network related to the political hashtag Zandberg.

Main Methods:

• Application of classical centrality measures (degree, betweenness, closeness) and local clustering coefficients.
• Community detection using the Louvain algorithm.
• Introduction and application of the Structural Entropy Index of a Community (SEIC).

Main Results:

• Empirical analysis of a real-world Twitter network around the political hashtag Zandberg.
• Significant variation observed in community structures and entropy levels.
• Larger communities exhibit decentralization (SEIC > 0.8), whereas smaller groups are often node-dominated.

Conclusions:

• The proposed methodological framework offers a robust tool for studying social network dynamics.
• Findings have implications for influencer identification and disinformation resilience.
• The SEIC metric provides a size-independent measure of communication decentralization.