In response to a recent collection budget cut and the expansion of the College of Engineering, the Engineering Library has initiated a comprehensive journal subscription assessment. While we have usage data on subscribed journals, we currently lack insights into the potential usage of non-subscribed journals. To address this gap, we leveraged faculty publications, references and grant information to assess journal subscription. Our goal of this lightning talk is to introduce a streamlined workflow and equip engineering librarians with practical tips to implement immediately.
The workflow of this project involves four steps.
1) Data retrieval: Faculty publication metadata including 1,944 journal articles and conferences published between 2019 and 2024 was provided in September 2024, by our colleague managing Esploro (our institutional repository and researcher profile). Over 95,000 reference metadata of faculty publications were retrieved from Scopus using an API with pybliometrics (Python-based API wrapper to access Scopus). Awarded grant information was provided by the College’s Research Support Manager.
2) Data processing: Normalizing messy journal titles in the metadata was challenging and common strategies such as lowering case were insufficient. To solve this, quality metadata of references were retrieved again from Scopus based on Scopus publication IDs. For grant information, keywords for each grant were generated by Copilot.
3) Data analysis and visualization: Journal titles were counted, and wordcloud for grant information were generated using R.
4) Human interpretation: Our graduate assistant checked selected journal titles against the library catalog system.
As a result, we identified a gap of 15 frequently used, non-subscribed journals, most of which align well with grant information. This project not only demonstrates the feasibility of performing an assessment within a limited timeframe but also provides actionable tips for others looking to undertake a similar approach