Finding where to contribute on Wikidata isn't always straightforward, especially with structured and lexicographical data spread across multiple WikiProjects and languages. Identifying a gap, and knowing how to act on it, can be harder than making the edit itself.
This session introduces Dustpan and Broomstick, two tools developed by Wikicollabs that help contributors discover opportunities to improve Wikidata. Dustpan helps contributors find Items with missing information by selecting a WikiProject, a category of missing information, and, where available, an additional scope. For example, contributors working with WikiProject Sum of All Paintings can use Dustpan to find paintings that are missing information, such as the creator or material used. Broomstick focuses on lexicographical data, helping contributors uncover Lexemes that can be improved across different languages, including issues involving Senses, Forms, and misplaced statements.
Through live demonstrations, participants will see how finding a gap leads to a specific edit they can make on Wikidata. By the end of the session, participants will be able to use Dustpan and Broomstick to find where to contribute on Wikidata. We also welcome suggestions for which WikiProjects and languages the tools should support next.
This session will provide an overview of Wiki Loves Names and its contribution to Wiktionary. It will highlight the importance of documenting personal names, their meanings, origins, pronunciations, linguistic forms, and cultural contexts, particularly names from underrepresented languages and communities.
The session will explore how Wiki Loves Names supports the preservation and visibility of naming traditions by encouraging contributors to add well-sourced information about names to Wiktionary. It will also highlight the value of Wiktionary as a free, multilingual dictionary and how name-related contributions can help improve knowledge about languages and cultures that are often underrepresented online.
The session will share the objectives, approach, progress, and potential impact of the initiative, while showcasing examples of contributions and opportunities for community participation. It will be an introductory overview rather than a hands-on training session.
For over 50 years, lawyers have used process models to describe how information and decisions move. Lexipedia, developed by the Center for Civic Innovation and the University of Virginia, describes legal processes using modern procedural standards such as BPMN (Business Process Model Notation) and Petri Nets.
Process models describe how things happen, but offer no convenient discovery mechanism. Lexipedia's latest release adds linking to pull Wikidata entries in directly, plus a QuickStatements generator for the Lexipedia Wikibase — where laws, case law, and other content not yet ready for the Wikidata ecosystem can await community review.
Linked data adds further value through specific legal ontologies: Akoma Ntoso, which correlates legal axioms across global legal systems, and deontic logic, which defines logic from language.
Cross-referencing legal concepts is self-explanatory; deontic logic is more nuanced. For example, if one MUST do something, it's implied there's a consequence for not doing it. In legal texts these statements often aren't adjacent, and get left out of models. Describing content with deontic logic lets us run a baseline "legal linting" pass that improves models and simplifies review.
See how the Georgia Tech Library designed and built a dataset of notable architecture alumni in Wikidata, for outreach and development. We will talk about the research need that led to the development of the dataset, reflect on the design and implementation challenges encountered along the way, and future plans for further development.
Place for linked data beginners, interested parties, and experts to gather to answer questions about linked data.
If you are currently working on a linked data project and are willing to answer linked data questions we could use the help! Let the organizer know when you arrive.
Designed for digital humanities research, Semantic Kompakkt is a free, open-source platform for exploring and semantically annotating visual media in an interconnected open-data environment. Semantic Kompakkt, provides accessible features through graphical interfaces while Wikibase manages metadata storage as linked open data, thus enabling the publication and annotation of 2D, 3D, and audiovisual media.
The proposed talk shares lessons from an evaluation of the metadata presented on Semantic Kompakkt field names as reviewed by digital archaeologists. The review was done as part of a project to reduce fuzziness within the platform. Fuzziness, also referred to as wobbliness, serves as an overarching concept encompassing indistinguishable combinations of uncertainty and vagueness within statements or datasets, as well as broader informational ambiguities such as unreproducible literature claims or imprecise ontological mappings.
As a result some key field names were updated, which include; the “persons and institutions” which became “actors”; and descriptions on object types, provenance, material, dimensions, collection and other such, which were re-directed to the original object data from the existing collection. This resulted in having to re-upload the items on Wikibase with all the new statements so as to better serve the objects on display in Semantic Kompakkt.
What does openness mean when data cannot leave a restricted environment? This lightning talk discusses how linked data principles are being adapted within the National Security Research Center (NSRC) at Los Alamos National Laboratory, where more than eighty years of secrecy have made access control a defining feature of knowledge organization.
At the NSRC, security requirements function as design parameters for semantic modeling, metadata standardization, and relationship building. These efforts support access and interoperability in an environment where data cannot be published on the open web. Here, openness takes shape through deliberate choices about what can be connected, who can follow those connections, and under what conditions.
Bridging library metadata to the semantic web sounds simple — until you're staring at a fuzzy title match and wondering whether "Emma" the Wikidata item is actually "Emma" the Jane Austen novel. This talk walks through a working pipeline that reconciles MARC records and spreadsheets against Wikidata before generating BIBFRAME RDF — deliberately in that order.
MARC's fixed fields, controlled identifiers, and $0/$1 authority links are a more reliable reconciliation substrate than post-conversion RDF would be: decades of cataloging discipline mean an ISBN or an LC name-authority code can be trusted at face value, in a way a freshly-generated triple can't yet be. Reconciling at this stage — while those signals are still intact — lets Wikidata's enormous, freely-reusable graph of works, people, and places get folded directly into the RDF as it's built, rather than bolted on after. MARC turns out to be less a legacy format to escape and more infrastructure in service of the linked data it produces.
The core mechanism is a tiered confidence system: identifiers auto-accept, fuzzy titles get a second look only when independently confirmed by author identity, and everything else is flagged rather than guessed at. I'll also cover verifying the BIBFRAME modeling itself against real production records — which caught errors theory alone missed.
Agentic AI coding tools such as Claude Code, OpenAI Codex, and the open source OpenCode enable users to create sophisticated software applications from natural-language feature descriptions. This approach, often called “vibe coding,” allows people without software development experience to build applications. It also changes how experienced developers can iteratively orchestrate AI agents to engineer software within highly compressed time frames.
This presentation describes an experiment conducted over a single weekend: developing a static-site RDF editor that uses a graph-based editing approach rather than the form-based interfaces used by tools such as Sinopia, Marva, and Share-VDE’s JCricket. The editor is available at https://ld4p.github.io/graph-editor/, and its source code is hosted at https://github.com/LD4P/graph-editor/.
Software Engineer, Stanford University Libraries, Software Developer
Jeremy Nelson is a software engineer at Stanford University Libraries. He is on the team that migrated Stanford to FOLIO and is active developer in the Sinopia Linked Data Editor and Blue Core projects. His research interests include applying AI to improving library workflows.
Tuesday September 29, 2026 3:20pm - 3:50pm EDT TBA
In this presentation, the PCC Sinopia Cataloging Affinity Group will share updates to the Sinopia editor made for Blue Core, including updates to the interface, functionality, and seed data. We will provide an overview of Sinopia, compare Sinopia Classic to Sinopia for Blue Core, and highlight plans for production cataloging in BIBFRAME. We will review recent developments to operational metadata such as data associated with the BIBFRAME AdminMetadata class, and outline ways to get started with BIBFRAME cataloging in Sinopia.
The LD4 Discovery Affinity Group was formed in 2019 to provide a space for discussion and sharing knowledge about the intersection of linked data and discovery. Recognizing the increasing focus or reference to AI in various library and discovery areas, this past year we started a series on AI, linked data, and discovery. We began by walking through fundamental concepts for machine learning and AI, and are working our way up to generative AI concepts, while also identifying overlap with linked data. In this LD4 conference session, in addition to providing a recap of the history of the group, we will also review topics and themes addressed in the series so far, and ask for input and suggestions regarding possible future topics.
