2019 Conference Objects and Reports

Using Probabilistic Topic Modeling of Library Access Records to Identify Learning Trends in Educational Research

Chen, Yi; Liu, Xiang; Chae, Hui Soo; Natriello, Gary J.

Advances in the architecture of digital library service infrastructure enable the collection of various types of data related to the use of library resources, tools, and services. The Big Data that is being generated provides valuable insight into library operations and has the potential to reshape the future of library work. In this paper, we describe the innovative application of topic modeling (supervised Latent Dirichlet Allocation) of research corpora accessed by patrons through a library proxy server. We found that the underlying topics of this corpus (e.g., psychology, family education, and methodology) converge with the general interests one would expect from a Graduate School of Education. In addition, we discuss the potential and challenges of utilizing library proxy log data in learning analytics research.