Library Textbook Collection Usage Patterns: Visualizing Library Usage Data
Spring 2009-Spring 2010
Table of Contents
Background
To help alleviate the costs of textbooks for students and their families, the North Carolina State University Libraries combined forces in 2009 with the University Bookstores to ensure that at least one copy of every required textbook for each class at NCSU is available on Reserve in the Libraries. The program has now been in place for three full semesters (Spring 2009 to Spring 2010). Before the start of each semester, the Bookstores provide the Libraries with a list of all the textbook adoptions from faculty on campus. If the books are already in the Libraries' collection, they are moved into a special collection within the library catalog called the "textbook" collection, and physically relocated to be held with reserve items behind the circulation desk. Items in the textbook collection can only be checked out for two-hour periods. Books that the Libraries does not already possess in its collections are ordered for the textbook collection. Currently, there are 3,921 items in the textbook collection.
Research Questions
- How has use of the collection changed over three semesters?
- Which books are used most frequently?
- What are trends in use over the semester and day?
Methodology
Circulation data from Spring 2009 to Spring 2010 for the textbook collection was exported from the Libraries' ILS for analysis. Data was analyzed using PHP and charts were created using Adobe Photoshop CS3 and Ggplot2: a freely available visualization package for the R software environment for statistical computing and graphics.
Several issues in data tracking arose that made some findings ambiguous. While data related to items in the textbook collection exists both in the Libraries' ILS and in spreadsheets kept by the NCSU bookstore, attempts to programmatically combine data from the two sources were futile due to lack of consistent unique identifiers across the data sources. Therefore, the data analysis conducted for this report was not able to take advantage of the data in the bookstore spreadsheets. While the ILS data made it possible to correlate a particular holdings item to its individual circulations as well as the date and time at which the circulation occurred, it was impossible to correlate all holdings items to the date they were added to the textbook collection. While the record creation date could be assumed to be the date that the item was added to the collection for items purchased specifically for the collection, a number of items already existed within the Libraries' collections and were simply reassigned to the textbook collection. There is no such data for these items. It was also impossible to correlate a particular holdings item to the semesters in which it was "active" (required for courses) using only the ILS data. A number of classes are offered only in the Spring or Fall semester, and some classes are offered every third or fourth semester. Without the ability to identify the semesters in which textbooks were active, how often an item has circulated can only be measured against the entire three semesters in which the textbook collection has been in place. This may create confusion regarding circulation numbers for all items that have either not been in the collection for the full three semesters or that have not been required for a class for all three semesters.
Because holds and requests cannot be put on items in the textbook collection, it was also impossible to see how frequently students wanted a textbook but could not have it because it was checked out. Such data would provide helpful information about which books are in the greatest demand.
Last updated: July 7, 2010