DEVCONNECT 2017: Dashboards and Artificial Intelligence in Libraries
Wayne State University Libraries has been working on several projects aimed at better understanding library material usage using OCLC products and APIs. The library staff’s efforts have led them to develop a library dashboard, a visualization tool that allows staff to see how users access materials via ILLiad interlibrary loan. This tool allows library staff to see not only usage, but also the costs associated with journal interlibrary loans. These costs include copyright clearance fees and vendor-supplied document delivery fees.
By understanding this data, staff can respond to fluctuations in journal demand and make collection management decisions based on usage changes. For example, a journal may incur multiple interlibrary loans leading to copyright clearance fees or multiple document delivery charges; in these cases, library staff can respond by reaching out to faculty to determine if this will be a sustained usage or by purchasing a new subscription. This tool uses the WorldCat Search API to standardize journal titles.
In addition to the library dashboard, Wayne State has been investigating artificial intelligence to better understand library collections and usage. One application still in development uses Library of Congress Classification Numbers to determine the likelihood that materials will checkout in the future. Using TensorFlow, an open-source artificial intelligence software package, the library fed in call numbers (standardized using the WorldCat Search API) and associated checkouts to a TensorFlow model aimed at predicting the likelihood an item will checkout. The resulting application combines a call number input via a webform with checkout history to determine how likely the item will checkout in the future.
This past May, I presented “Dashboards and Artificial Intelligence in Libraries” at the first annual DEVCONNECT conference hosted by OCLC. The video below outlines how Wayne State used OCLC products and APIs to develop a statistical dashboard to forecast journal article demand and used an artificial intelligence application to understand physical material usage in the library. DEVCONNECT was a great opportunity for information professionals from around the world to share how they have been developing software using OCLC products.
Assistant Director for Discovery and Innovation, Wayne State University Libraries