Spectral methods for detecting periodicity in library circulation data: a case study

Abstract

The purpose of this investigation is to show the feasibility of spectral methods in the field of information science, in particular for the analysis of library circulation data. Using the software package MATLAB® we applied the discrete Fourier transform to obtain frequency information about the noisy time series of circulation data. Over a time span of two academic years we could clearly detect a semestral and—less visibly—a weekly periodicity. Other periods were not so distinct and could be spurious. The normal loan period of four weeks could not be detected. Following McGrath (1996, Journal of the American Society for Information Science, 47, 136–145), and Naylor and Walsh (1994, Library and Information Science Research, 16, 299–314) we conclude that spectral methods show a lot of promise for analyzing all kinds of time series and other signals occurring in the field of information science. This approach certainly deserves more attention from practitioners.