Publishing your EM dataset to Zenodo

Zenodo is a general-purpose research data repository operated by CERN, supporting versioned datasets with persistent DOIs (Digital Object Identifiers). It’s the canonical publication path for Extended Matrix-based reconstructive datasets that need to be cited, archived, and made openly accessible.

This recipe sketches the workflow and the caveats. The page will be extended with a worked example (the Great Temple of Sarmizegetusa project, currently in preparation) as soon as that publication goes live.

ORCID — get one first

Before publishing anything on Zenodo, get an ORCID iD if you don’t have one. ORCID is a persistent researcher identifier that links your publications, datasets, and software contributions across platforms. Zenodo integrates with ORCID natively. Beyond Zenodo, the ORCID iD is recommended throughout the Extended Matrix ecosystem as the canonical author identifier — it makes paradata authorship traceable when projects are forked, merged, or shared across teams.

What a DOI gives you

When you publish a deposit on Zenodo, the platform mints a DOI that points uniquely and permanently to that specific version of your dataset. The DOI:

  • Survives institutional or project domain changes.

  • Lets others cite the exact dataset you published, not a moving target.

  • Is indexed by scholarly search engines and citation databases.

  • Allows reuse under the license you assign (CC-BY, CC0, etc.).

In practice, a DOI’d dataset is a real publication. Treat it like one when you prepare it: clear authorship, clear license, clear description, complete files.

Rights — verify before publishing

Before publishing anything in your DosCo or any other EM artefact, you must be certain you have the right to publish each individual item. This is the most error-prone part of the workflow.

Particular attention to:

  • Extractor nodes and their attached media are usually intellectual property of the authors of the extracting work — they are derivative works (an annotation, a quotation, a measurement) layered on top of an original source. The authors of the extractor (you, your team, third-party collaborators) hold rights on the extraction; the underlying source may belong to someone else. Verify case by case.

  • Original sources (historical photographs, archival documents, archaeological survey deliverables, prior publications) may carry separate copyrights. Some are in the public domain; others require permission to redistribute.

  • 3D survey models may belong to the surveying team or to the contracting institution — not automatically to you.

  • Reconstructive proxies and representation models you authored in EM Tools are your own work and are publishable under your chosen license, unless they incorporate textures or geometries with third-party rights.

When in doubt about a single document or media file in your DosCo: do not include it in the Zenodo deposit. Instead, publish a pointer (URL, archive reference, or bibliographic citation) to the document where it lives online, and let the graph reference that pointer rather than carrying the file.

Folder structure inside the deposit

Use the canonical Extended Matrix folder structure inside your Zenodo deposit. The same tree that organises your local workspace (DosCo/, matrix/, model/, source_list.xlsx, etc. — see EM workspace preparation) carries to the deposit. This makes the dataset legible to anyone familiar with EM, and lets downstream tools (custom mappers, future bulk-import scripts) parse the deposit without bespoke unpacking logic.

Versioning

Zenodo supports versioned deposits: when your dataset evolves (new data, corrections, fresh interpretations), you publish a new version with its own DOI under the same conceptual identifier. The conceptual DOI points to the latest version; each version DOI points to that specific snapshot.

Plan your versioning strategy at the start:

  • v1.0.0 at first publication, even if the work continues.

  • Bump the version for substantive changes that affect the reconstruction or the evidence base.

  • Use the Zenodo description field to summarise what changed between versions.

Immutability — once published, the files cannot be deleted

Zenodo applies a SHA digital signature to every published file. Once a version is live, the files are immutable: they cannot be deleted, edited, or replaced. You can publish a new version of the deposit, but the previous version’s files remain available at their original DOI.

This is a feature, not a limitation — it’s the foundation of citability. But it also means:

  • Sensitive or rights-questionable files must not be in the deposit in the first place (see Rights, above).

  • Personal data, anything that could violate GDPR if revealed, must be filtered out before upload.

  • Pre-publication review (by you, your co-authors, your institution) must happen before the “publish” button is pressed.

Treat the deposit pre-upload as you would treat a journal manuscript proof: assume everything you put in is going to be permanent.

Worked example

The first published Extended Matrix dataset under this workflow will be the Great Temple of Sarmizegetusa project — currently in preparation. This section will be expanded with the concrete deposit metadata, file inventory, and lessons learned as soon as that publication goes live.

Future direction

A direct exporter from EM Tools to a Zenodo-ready deposit bundle is on the development tracker (see the development projects for status). Until that ships, the upload to Zenodo is a manual operation: prepare the folder, zip if needed, upload via the Zenodo web interface, fill in the metadata, attach the ORCID, set the license, mint the DOI.