X as a Passive Sensor to Identify Opinion Leaders: A Novel Method for Balancing Visibility and Community Engagement

Abstract

The identification of opinion leaders is a matter of great significance for companies and authorities, as these individuals are able to shape the opinions and attitudes of entire societies. In this paper, we consider X (formerly Twitter) as a passive sensor to identify opinion leaders. Given the unreliability of the traditional follower count metric due to the presence of fake accounts and farm bots, our approach combines the measures of visibility and community engagement to identify these influential individuals. Through an experimental evaluation involving approximately 4 million tweets, we showed two important findings: (i) relying solely on follower count or post frequency is inadequate for accurately identifying opinion leaders, (ii) opinion leaders are able to build community and gain visibility around specific themes. The results showed the benefits of using X as a passive sensor to identify opinion leaders, as the proposed method offers substantial advantages for those who are involved in social media communication strategies, including political campaigns, brand monitoring, and policymaking.

Cryptocurrency Turmoil: Unraveling the Collapse of a Unified Stablecoin (USTC) through Twitter as a Passive Sensor

“…In the realm of opinion leader analysis [ 8 , 9 , 10 ], diverse methodologies have been proposed to identify opinion leaders. Typically, these approaches involve a blend of visibility and community engagement metrics, encompassing factors like the number of followers, followings, likes, replies, quotes, and mentions.…”

“…Generally, these individuals wield significant influence on attitudes, beliefs, and behaviors [ 22 ]. In the social media context, opinion leaders are perceived as field experts, and recognizing them is crucial for enhancing recommendation accuracy and the efficacy of social media campaigns [ 8 ].…”

“…The conventional method for pinpointing opinion leaders often involves associating them with the accounts boasting the highest follower counts. However, this approach is inherently flawed for two primary reasons: firstly, contemporary content algorithms on social media platforms prioritize discovery, granting smaller creators the potential for substantial reach and influence; secondly, the count of followers can be artificially inflated by fake accounts and bots, rendering it an unreliable metric [ 8 ]. Table 2 outlines the top 15 accounts within our dataset based on the follower count.…”