Appendix C — LīvMDb: Livonian Music Database

The Livonian Music Database (LīvMDb) is a proof-of concept for working with music that has low documentation depth, weak institutions. The music of the Livonian people is scattered, and as native speakers of this small ethnic group died out, their heritage was dispersed, and largely not placed on modern digital platforms. The LīvMDb as a functional module of the Open Music Observatory, serves as a testbed for working with very low documentation, decolonisation, and other issues related to regional cultures, ethnic minorities. It offers insights into subsidiarity.

The LīvMDb is federated with the Finno-Ugric Data Sharing Space and the Open Music Observatory. The first places Livonian music into a wider Finno-Ugric cultural context, the latter into a music context.

The LīvMDb is supported by a data-sharing space consisting of both shared databases.

The data-sharing space currently comprises the following initial components:

The Livonian Metadata Database serves as a support layer that is partly public and partly private. Its metadata definitions and descriptive metadata are exported into the LīvMDb databases as needed and permitted.

The Livonian Metadata Database

The Livonian Metadata Database is developed in alignment with the metadata framework of the Open Music Observatory.

  • Ontological and thesauri patterns: Reuses standardized or widely adopted vocabularies.

  • Conceptualizations and definitions: Includes concept definitions, thesauri, and other elements developed specifically for the LīvMDb.

  • Public permanent identifiers: Uses identifiers that are public or can be made public.

The metadata layer is generally licensed under CC0, though in some cases other licenses are used (for example, CC-BY).

Note

Examples:

  • The definitions of musical work, music recording and is recording of relationship allow the description of connections between an abstract musical work and its recording(s).

  • The ISRC code identifies recordings across all streaming platforms.

Livonian Music Database (public)

The Livonian Music Database is a linked open database published by the Reprex on behalf of the Open Music Observatory. The database is distributed under various Creative Commons licenses that allow both commercial and non-profit use.

The primary aim of the LīvMDb is to provide a use case for very low documentation music ecosystems with very limited resources and challenging data curatorial scenarios.

Livonian Music Database (private)

The private components of the LīvMDb are a staging area for data that has unclear provenance or legal status. Unlike in the case of LīvMDb, we hold minimal business confidential data (related to the royalty accounts of music that we published), but some data may have, for example, unclear GDPR status.

C.0.1 Microdata

Microdata consists of information before it is aggregated into statistical datasets or formal publications. Metadata can also be considered microdata: while it is never aggregated, it plays a critical role in describing the provenance, semantics, and usability of aggregated data.

  • Collections: Structured sets of similar items created through curatorial activities, where inclusion is based on discretionary selection to serve end users (e.g., a library’s holdings or a curated playlist). Our work is centerred around the work of Hõimulõimed, a Finno-Ugric NGO, which curated many Finno-Ugric language collections, including the collection of Livonian-language songs available on Spotify.

  • Registers: Authoritative lists created through administrative processes with defined rules, aiming to capture all known items in a category. Our work buids on the register of Livonian placenames, because they offer the most straighforward curatorial help to find new Livonian (folk) music1.

  • Metadata: Relevant elements from the Livonian Metadata Database that support the use of collections or registers.

The LīvMDb’s collections and register datasets are organized as a document database. This database stores structured data in RDF format describing musical works, sound recordings, printed and manuscript scores, as well as biographical information about music professionals and their organizations.

Each music-related object or agent (person, corporate body, or organization) is represented as a microdata dataset. These datasets share common definitions via conceptual models and data structures, enabling automatic aggregation. All datasets are available with RDF annotation and can be exported in all standard RDF serializations.

Microdata is intended for institutional and professional use, not for the general public. It is annotated with standardized metadata suitable for applications such as music library cataloguing, distribution platforms, and rights management systems.

Note

Example

The village of Mazirbe (Livonian: Irē, German: Klein-Irben, Russian: Мазирбе) is the central location of the Livonian culture, which hosts the Livonian Community House. Folk songs collected and recordings made in Mazirbe may have a provenance of Irē, Mazirbe, Klein-Irben (or its Finnish and Estonian versions), or various Cyrillic transliterations. Mainitaing a clear metadata dataset on the geographical sources of Livonian music is necessary.

  • Graphical view: Navigate and contribute to detailed entries on musical works, sound recordings, and related assets. 👉 Explore on Wikibase

  • Semantic view: Export structured data in XML, JSON-LD, Turtle, or N-Triples formats for reuse in research or digital projects. 👉 Example Turtle file or download XML

We made available the dataset in standard RDFXML, JSON-LD, and TTL serialisations accompanies by a data paper explaining its use (Antal et al. 2025; Antal, Pigozne, and Mester 2025)

C.0.2 Statistical Data & Data Catalogue

Our statistical data and catalogue consist of datasets aggregated using statistical methodologies. These comply with the SDMX standard and the W3C Data Cube vocabulary, making them compatible with spreadsheet software, statistical packages, and data science workflows in R, Python, or similar environments.

Datasets are offered in multiple formats. In addition to RDF serializations, we provide standard CSV files and, when required, Excel or SPSS formats. We also publish data papers and related documentation that describe dataset usability and highlight key insights.

C.0.3 Publications & Catalogue

The LīvMDb’s most important publications are sound recordings, which are made available as a archival recordings (not ready to be communicated to the public on commercial platforms) and publicly available sound recordings. We also provide access to some printed and and hand-written scores.

Microdata and statistical datasets are treated as publications and are listed both in the general catalogue and a machine-readable data catalogue.

The LīvMDb also includes methodological and musicological publications, as well as data papers explaining the use of datasets. The catalogue is designed for interoperability with libraries, archives, museums, and similar institutions.

  1. Livonian place names: documentation, problems, and opportunities (Ernštreits 2020) and our gazetteer dataset: (Antal, Pigozne, and Mester 2025).↩︎