Modeling time-to-trigger in library demand-driven acquisitions via survival analysis

Conventional statistical methods (e.g. logistics regression, decision tree, etc.) have been used to analyze library demand-driven acquisitions (DDA) data. However, these methods are not well-suited to predict when acquisitions will be triggered or how long e-books will remain unused. Survival analysis, a statistical method commonly used in clinical research and medical trials, was employed to predict the time-to-trigger for DDA purchases within the context of a large research university library. By predicting which e-books will be triggered (i.e., purchased), as well as the time to trigger occurrence, the method tested in this study provides libraries a deeper understanding of factors influencing their DDA purchasing patterns. This understanding will help libraries optimize their DDA profile management and DDA budgets. This research provides a demonstration of how data science techniques can be of value for the library environment.