Text Mining-Based Analysis of Content Topics and User Engagement in University Social Media

Abstract

A large amount of text data flows through the web every day in the context of various comments and reactions of users to social network content. Such information tokens can act as quality metrics for content, a group, or an industry. This information can be valuable for decision makers and content managers of higher education institutions (HEIs) as well as researchers studying the marketing activities of HEIs. Topic modelling, as one of the methods for analysing textual information, can help practitioners and researchers create analytical models that allow them to make important management decisions based on audience reactions to content. As part of this task, the goal of this research is to develop a topic model based on LDA to uncover key topics of posts in 15 university groups on the “VK” social network of five leading universities located in Saint Petersburg, Russia. For each university, posts from three groups were collected: “Overhead,” “Main” and “Enrolment.” In total, 247,633 posts were collected over the 2018–2023 period, resulting in more than 6 million tokens. During the application of the methodology, large clusters of posts on social networks were identified. Using analytics of reaction metrics and feedback from subscribers of university groups, recommendations were made for managing community content, and a convenient decision-making matrix was developed based on the bubble “Comments-Likes” diagram. These results can be used by decision makers to improve university brands and increase engagement in social networks.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 2 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.