Small Number of Communities in Twitter Keyword Networks

TL;DR

This study explores the community structure of Twitter keyword networks, revealing they tend to form a small number of communities, supported by empirical data and AI-generated pseudo-tweets.

Contribution

It introduces the hypothesis that Twitter keyword networks have few communities and validates this through analysis of real and AI-generated tweet data.

Findings

Twitter keyword networks exhibit a small number of communities

Empirical data from political figures supports the hypothesis

AI-generated pseudo-tweets reinforce the small community structure

Abstract

We investigate networks formed by keywords in tweets and study their community structure. Based on datasets of tweets mined from over seven hundred political figures in the U.S. and Canada, we hypothesize that such Twitter keyword networks exhibit a small number of communities. Our results are further reinforced by considering via so-called pseudo-tweets generated randomly and using AI-based language generation software. We speculate as to the possible origins of the small community hypothesis and further attempts at validating it.