Identifying Distinct Borrower Profiles in University Libraries Using K-means Clustering: A Case Study of Songkhla Rajabhat University

Abstract

University libraries generate vast amounts of borrowing data that remain underutilized for strategic decisionmaking. Traditional statistical approaches fail to identify hidden patterns in complex, multidimensional library usage data. Libraries lack systematic methods to segment users and optimize resource allocation based on empirical borrowing behavior patterns rather than assumptions. This study applies K-means clustering to $\mathbf{2 6, 8 5 7}$ borrowing records from Songkhla Rajabhat University (2021-2024), transforming raw data into 22dimensional feature vectors including temporal patterns, book categories, frequency, and demographics. The Elbow Method determined optimal four clusters. Four distinct borrower segments were identified: Course-Specific Management Students (February peak, 39.8 % management books), Diverse Academic Borrowers (consistent year-round activity, 15.6 avg books), Summer Vacation Borrowers (July focus, 79.9 % peak), and New Academic Year Borrowers (August-October concentration). The algorithm achieved a silhouette score of 0.72, indicating wellseparated clusters. This research provides a reproducible methodology for libraries to implement data-driven user segmentation, enabling targeted collection development and temporally optimized resource allocation. The approach transforms unstructured borrowing data into actionable insights for evidence-based library management.

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 — 1 citations exist upstream.

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