Browsing by Subject "R"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    A Method for Verifying Indicators of Journal Quality
    (2018-10-26) Odell, Jere D.; Polley, David E.
    A recent search of the UlrichsWeb Global Serials Directory for active, digital, peer reviewed, scholarly journals shows that world’s academic articles are published in more than 58,500 journals. By one estimate the growth of new journal titles increases by 2.5% ever year (Ware & Mabe, 2015). At the same time, universities are adopting researcher information systems that provide administrators and other campus stakeholders with nearly complete bibliographic data for all articles published by their faculty authors. As campus leaders work to make sense of this data, they may turn to their library for help. Questions may include: Are all of these new or previously unencountered journal titles legitimate? Who are the main publishers of our articles? What are the emerging trends that promotion and tenure committees should consider? The most common way to address these questions involves significant shortcomings--proprietary subscription databases, like Scopus, Web of Science, and Academic Analytics, have limited coverage of the journal literature and, by design, are unlikely to include newer and lesser known journal titles. At the same time many universities publish thousands of articles per year, manually checking each article submitted to a faculty annual review database would prove to be a tedious and lengthy process. To reduce the labor involved in identifying indicators of journal quality, we have developed a method using open source software and open Application Programming Interfaces (APIs). In specific, our method reduces the labor in identifying the publishers for a long list of journals and in identifying the access model for these journals (subscription-only or open access). To do this we wrote an R script that uses the SHERPA RoMEO and the DOAJ APIs. Using this method permitted us to quickly identify the journals that needed closer inspection. This method will help others that are working to verify journal quality in large data sets without relying on problematic, journal blacklists.
  • Thumbnail Image
    Item
    Using the ‘rentrez’ R Package to Identify Repository Records for NCBI LinkOut
    (code4lib, 2017-10-18) Lee, Yoo Young; Foster, Erin D.; Polley, David E.; Odell, Jere D.; University Library
    In this article, we provide a brief overview of the National Center for Biotechnology Information (NCBI) LinkOut service for institutional repositories, a service that allows links from the PubMed database to full-text versions of articles in participating institutional repositories (IRs). We discuss the criteria for participation in NCBI LinkOut for IRs, current methods for participating, and outline our solution for automating the identification of eligible articles in a repository using R and the ‘rentrez’ package. Using our solution, we quickly processed 4,400 open access items from our repository, identified the 557 eligible records, and sent them to the NLM. Direct linking from PubMed resulted in a 17% increase in web traffic.