Ecological metrics of diversity in understanding social media

TL;DR

This paper explores the application of ecological diversity metrics to topical discussion networks on social media, assessing their usefulness in understanding network composition and diversity.

Contribution

It introduces the analogy between social media discussion networks and ecological populations, evaluating three ecological diversity metrics for their applicability.

Findings

Each metric provides unique insights into network diversity.

The choice of metric depends on the specific research question.

Ecological metrics can be valuable tools for analyzing social media discussions.

Abstract

Topical discussion networks (TDNs) are networks centered around a discourse concerning a particular concept, whether in real life or online. This paper analogises the population of such networks to populations encountered in mathematical ecology, and seeks to evaluate whether three metrics of diversity used in ecology - Shannon's $H'$, Simpson's $\lambda$ and $E_{var}$ proposed by Smith and Wilson - give valuable information about the composition and diversity of TDNs. It concludes that each metric has its particular use, and the choice of metric is best understood in the context of the particular research question.