Seeing the Forest by Counting the Trees

Abstract

Libraries, particularly academic libraries, are swimming in a sea of data. Librarians often contribute to this by counting every possible patron interaction in an attempt to both define their current situation and to predict future staffing, budgetary, and collection needs. This investigation assessed the effectiveness of using various data sources in predicting future library activity and needs. The authors collected data on in-person and chat reference transactions, electronic journal downloads, database queries, and catalog searches from 2009–12. By analyzing these data points, the authors hypothesized they would find correlations that might be predictive of changes in related library services. Results indicated that the strongest correlations track activity over the course of the academic calendar. While none of the data points examined had predictive properties, the strong correlations between the data points over the period of time studied indicated that any one of them might serve as a stand-alone indicator of usage.