Final Report --- Always Already Computational: Collections as Data

2020 NDSA Agenda

“…This work includes building functional requirements such as support for collections-as-data infrastructure development, developing use cases, and producing a collections-as-data framework that will ultimately encourage the computational use of digitized and born digital collections. 56 Experiments around the research use of web archives are instructive to stewards of all types of digital content. The Archives Unleashed project, 57 for example, is an effort to make petabytes of web archives data accessible to researchers for scholarly analysis, while also providing the tools and cloud computing environment necessary to do so.…”

Reusing digital collections from GLAM institutions

“…The FAIR data principles provide guidelines to improve the findability, accessibility, interoperability and reuse of digital collections [12]. Collections as data provide best practices for turning digital collections into data sets amenable to computational use and novel research methods [5]. As a result, many organisations recommend publishing machine actionable collections, combined with documentation about digital collections [13,14].…”

“…In this sense, new approaches have recently appeared such as the concept of Labs within GLAM institutions that facilitates the adoption of innovative and creative tools for content delivery and user engagement [4]. In addition, new methods seek to address the publication of digital collections as data sets amenable to computational use such as collections as data [5]. Making digital collections available as data and ready for computational analysis have an impact on user engagement since collections are more accessible and interoperable, using open licences and reactivating legacy material held by GLAM institutions.…”

“…Institutions such as the World Wide Web Consortium (W3C) recommends the use of machine-readable standardised data formats, providing data in multiple formats and reusing vocabularies [5,45]. Making data available in machinereadable relates to the use of standardised data formats that are easily parsable including but not limited to CSV, XML, JSON and RDF.…”

A benchmark of Spanish language datasets for computationally driven research

“…12 New approaches such as Collections as Data provides a framework to create machine-actionable collections ready for reuse. 13 The Library of Congress (LC) recommends creating digital collections usable for computation as well as building institutional capacity for digital scholarship and for expanding user services. 14 The Online Computer Library Center (OCLC) recently published a study on community engagement with data science, machine learning, and artificial intelligence.…”