Application of Topic Modeling and Data Visualization in Libraries: A Review of Literature 8 7

Abstract

This study explores the evolving role of topic modeling and data visualization within Library and Information Science (LIS), focusing on their theoretical underpinnings, practical applications, and recent research trends. To clearly understand the techniques, 73 scholarly publications, published between 2003 24, were retrieved from different databases such as Scopus, Web of Science, LISA, and Google Scholar. A detailed review of the literature of the publications was done to understand the theoretical aspects of the techniques and their prospects. Topic modeling is a sophisticated text-mining technique that can be done using a variety of algorithms such as LDA, STM etc., each one suited to different types of datasets and research objectives, with diverse applications in libraries such as Enhancing Information Retrieval and Search Systems, Improving Digital Library Services, Document Clustering and Recommendation Systems, Subject Metadata and Multimedia Libraries, Topic Trends and Knowledge Discovery in LIS etc. Analytical findings of the publications reveal that 75% of the literature emerged in the past decade, with a notable concentration between 2018 and 2022, indicating heightened academic interest in these data-driven methodologies. The reviewed documents include journal articles (70%), conference papers (18%), book chapters (9%), and theses (3%), highlighting diverse scholarly contributions. Topic modeling techniques, particularly Latent Dirichlet Allocation (LDA), are recognized for uncovering hidden patterns in large text corpora, while data visualization translates complex datasets into impactful visuals. Together, these tools offer immense value to LIS professionals, data analysts, and researchers seeking to advance knowledge discovery, improve decision-making, and enhance scholarly communication in modern libraries