Data-driven collection development: Text mining college course catalogs

Academic librarians are trained to develop and manage collections. They rely on their own subject expertise in academic disciplines, input from teaching faculty, and professional training to make informed selections to support institutional curriculum. Professional training in collection development has, in recent years, focused on evidence-based acquisition methods (Johnson, 2018, p. 134). College and university course catalogs are a potential but untapped source of evidence for identifying topics of importance to institutional curricula. Course descriptions are concise descriptions of the subjects covered in college or university courses and therefore the topics about which students may require additional sources of information. Until recently, examining course catalogs was a time-consuming prospect. The advent of data and text mining techniques, however, makes it possible to analyze course descriptions with much less time and effort expended. This article contains a brief introduction to data science in libraries; details of tools and processes used for collecting and cleaning course catalog data; and preliminary results of a project to mine course catalogs for changes in curriculum focus to benefit library collection development decisions.