Best Practice

Datafication is a transformative process that not only converts data production into machine-readable format, but also renders it machine-processable and actionable. This study centers on a case study of the state investigations of popular religious sects during the early 19th century of the Qing dynasty in China. In the field of historical research, official documents have emerged as vital subjects and data sources for digital humanities research, owing to their capacity to furnish concise or elaborate information on diverse historical, political, and social events. However, utilizing these documents as research data sources presents challenges due to the lack of digital text conversion, lack of structures and punctuations given their antiquity, and incomplete information pertaining to specific entities such as individuals, organizations, and events is dispersed across disparate document texts. To tackle these challenges, experts in historical research, library science, and information engineering collaborated to develop strategies for transforming historical archives of official documents into digital data. This project developed digital research methodologies and automated tools, including text analysis and Optical Character Recognition (OCR), to facilitate a human-machine collaboration aimed at extracting pertinent and significant entity information germane to the exploration of Qing official archival records pertaining to popular religions. A domain ontologies was created based on the CIDOC CRM (Conceptual Reference Model) while extending the TEI (Text Encoding Initiatives) framework to contextualize and structure the extracted entity information derived from the archival content. The project subsequently developed best practices encompassing linked data conversion, SPARQL templates for question answering, and visualization techniques, which collectively showcase the pivotal role of datafication within the context of digital humanities research. The methods developed from this project can foster collaboration between humans and machines through the utilization of Generative AI tools, such as ChatGPT.

Data-literature linking is a necessity for the sharing and reuse of scholarly resources, and accelerating scientific progress, current linking solutions are primarily constrained in the dimension and depth of relationships between research data and scientific literature. The presentation outlines a framework for data-literature linking that combines data enrichment techniques such as automatic discipline classification, subject indexing, and entity extraction based on semantic knowledge bases. Furthermore, a prototype system for data-literature link discovery is introduced, which includes research data retrieval and faceted browsing, data-literature and data-data bidirectional relation discovery, and entity-based link browsing and navigation.

One of the key bottlenecks in international development is the appraisal and approval of projects. This involves the documentation of project development activities prior to implementation. In developing countries, the size of projects varies widely, but the project budget is often smaller scale than in developed countries and the needs are greater. This adds up to more projects with smaller budgets that need to be on-boarded and managed. This is a major documentation challenge. This presentation discusses the metadata strategy of SANKOFA, the African Development Bank’s latest document and records management platform that aims to provide the documentation facility to address this challenge. This presentation describes the design and definition of metadata, and strategies for automating generation of high-quality granular metadata. In document and records management projects like SANKOFA, metadata specification, collection and use rely on developing and maintaining an organization-wide multi-level file plan. We will discuss best practices for file plan and metadata design, validation, implementation, and maintenance in a multilateral financial institution setting.

The library domain has been actively seeking ways to transition towards metadata creation and discovery in linked data using semantic web technologies and the BIBFRAME, a library-specific ontology. While libraries have made significant progress in ontology and prototype developments, there are still critical areas that require attention, particularly regarding the initiation of library data creation in linked data within the library's integrated system. This research aims to provide an overview of the current status of library linked data development and present specific recommendations on the essential features, ontologies, and services that should be considered when developing an integrated linked data creation system that supports linked data-based discovery services. (*A part of this research was presented at the Alma Linked Open Data Community Meeting on April 26, 2023.)

The Staatsbibliothek zu Berlin has been developing and operating the CrossAsia platform (https://crossasia.org) for more than a decade. It was launched in 2005 to provide easy-to-use and central access to the library's printed and electronic resources on East, Central and Southeast Asia. The collection is partly funded by the German Research Foundation (DFG). Over the years, the Staatsbibliothek zu Berlin has continuously adapted the objectives of CrossAsia, taking into account and in part anticipating the needs of academia and research. In this short overview, I would like to briefly highlight the various challenges in dealing with data from Asia (China, Japan, Korea, Taiwan, Mongolia, etc.) in a German library environment. I will present our ideas and strategies for dealing with mostly proprietary (and non-Latin script) data and our approaches to providing standardised and easy-to-use services, such as our discovery based on VufFind.

Metadata enrichment is a way to augment the description of cultural heritage entities, improving data discoverability and supporting end-users in having access to critical context for a given entity. In the Europeana network, several aggregators and projects aim to enrich metadata for cultural heritage objects that they provide. A variety of processes are used to produce enrichments, from fully manual (e.g., using crowdsourcing) to fully automatic (e.g., geo-enrichment), and these vary in terms of quality, reliability and informational value. Therefore, before the enrichments can be integrated into the Europeana services, they must be validated. Europeana is defining a general methodology to assess the overall quality of the enrichments produced by a project or a tool. This presentation will introduce the methodology and describe the outcomes of its first applications in the projects Jewish History Tours and EnrichEuropeana+.

This work has been carried out with Valentine Charles, José Eduardo Grano de Oro, Nuno Freire, Lianne Heslinga, Hugo Manguinhas, Monica Marrero, Ilektra Osmani, Ad Pollé, Paolo Scalia, Eleftheria Tsoupra (Europeana Foundation); Alexander Raginsky, Vadim Shestopalov (Jewish Heritage Network); Sergiu Gordea, Lukas Mölschl, Michela Vignoli, Medina Andreșel, Srdjan Stevanetic (AIT Austrian Institute of Technology); Frank Drauschke, Jasmin Salesevic (Facts & Files Berlin); Łukasz Krzyszczuk, Adam Poznański (Uniwersytet Wrocławski); Sonja Galina (State Archives Zagreb)

The National Diet Library (NDL), in cooperation with academic libraries and research institutions in Japan, has developed and published the Metadata Distribution Guidelines in the hope of improving interoperability of metadata to digital resources. The Metadata Distribution Guidelines were written to help ensure interoperability of metadata via systems for promoting the utilization of digital resources like NDL Search and IRDB. The Guidelines provide stakeholders involved in the design, creation, and maintenance of metadata with a means for determining an appropriate design and distribution of their metadata. The Guidelines provide information corresponding to metadata characteristics of material forms, such as research data and pre-modern materials, in addition to information common to material formats. The dissemination of the Guidelines is intended by the NDL to improve the discoverability of content by reducing the cost to individual institutions of independent examination for metadata interoperability as well as to establish an efficient and sustainable metadata distribution channel. The NDL also expects the Guidelines to function as a communication tool for metadata distribution.

The concepts, guidelines, and best practices related to metadata are well-known factors in the global context. Hence, the weight given to this domain is varied in each country. The study investigates some fundamental conceptions of metadata and related standards through various literature and presents some best practices of metadata related to Brazil and Sri Lanka. Both countries have some initiatives related to library and geospatial data domains. The scale of these projects may be different, but we can see some similarities in both scenarios. Basically, libraries use various metadata standards to organize and retrieve resources, and this applies to both countries. Compared to Brazil, the Sri Lankan Library, Archive, and Museum (LAM) awareness of metadata is confined to MARC standards. Similar to Brazil, many institutions in Sri Lanka are maintaining Dspace repositories that use a qualified Dublin Core-based metadata schema. Some professionals in the information science sector are aware of Dublin Core standards, but the use cases are very rare. Based on the above best practices, awareness of Dublin Core metadata standards in Brazil is wider compared to Sri Lanka. However, this scenario should be further investigated thoroughly. Finally, awareness of basic conceptions and standards related to metadata is a key factor when it comes to conducting more research in the domain. Currently, LIS students are taught about metadata standards during their postgraduate programs, and regular webinars and workshops are conducted by universities and institutions such as the National Library of Sri Lanka and Brazil to make people aware of the use of metadata in diverse domains.

Every (meta-)data based project has some form of quality checks. At Europeana, we begin with checking the syntactic validity of the metadata we receive, against the specification of our data model. However, this is not enough, and a "long tail" of issues like incomplete information, lack of normalisation and misuse of fields still have a negative impact on relevancy for the user, for example when search results and item pages are cluttered with not-so-relevant information. To minimize, if not eliminate, such data quality problems in Europeana, the Europeana Data Quality Committee has been looking for ways to automatically identify these problems, so they can be flagged to Europeana’s data partners early in the data ingestion process and if possible solved. This presentation will provide a brief overview of established data quality problems that occur in the Europeana database and describe our approach so far to address these problems using Europeana’s Metis Sandbox (a test environment for Europeana’s data) as a detection and reporting mechanism.

Since 1998 Korean academic institutions have shared cataloging in MARC (Machine Readable Catalog) format and over 800 institutions are joining now. Most of the new records come from university libraries and metadata is mainly MARC driven. Among 13 million records 17% of them were issued in linked open data and the most instances are from thesis and dissertations. Book records of one university were issued in linked open data at once. The RISS(Research Information Service System) our organization(KERIS, Korea Education and Research Information Service) runs has various materials in one database, but only the records in MARC format can be transformed to the linked open data. Many researchers are interested in current research outputs and those materials are integrated with the research portal system such as PRISM, POINT, and NKIS. However, data gathering and interoperability are for searching purpose only, not out-bounding the contents to the portal. We will show the instances in different materials and how they can extend the usage and authority profile. In addition, we will make RISS research paper records to the linked open data to refine the usability.

Ancient Chinese Inscription Rubbings are one of the carriers for the inheritance of Chinese history and culture. In addition to their literary value, they also have extremely high artistic appreciation value. In order to better reveal the historical, cultural, and artistic attributes of the Ancient Chinese Inscription Rubbings resources, we have designed an ontology application profile that comprehensively and meticulously describes the literary attributes such as title,creator,engraver,material,size,etc.,artistic attributes like style of calligraphy, and the craft and techniques in the transformation process from stone tablet to rubbings , as well as the history of being collected and appreciated by ancient epigraphers. At the same time, we have refined the granularity of resource description to individual characters in the rubbings, including the position, order, font, and font shape of individual characters in the original rubbings, to help researchers understand the cultural history carried by the inscription rubbings from different perspectives, compare the artistic styles of different artists, and better assist calligraphy enthusiasts in learning and appreciating Chinese calligraphy art. Based on this Ontology AP, a Knowledge Base was built and it provides bibliography retrieval, historical and cultural Exhibition,and art appreciation services in the new Shanghai Library in Pudong.