Skip to content

🌐 Add Your Data to the MatWerk Knowledge Graph (MatWerk KG)

The MatWerk Knowledge Graph (MatWerk KG) improves data visibility and serves as an indexing and discovery layer for materials science and engineering resources. It helps people find, connect, and reuse data more easily.

Adding more data benefits us all. Depending on your data type and how your data is currently stored, choose the scenario below that best matches your case.

🚀 Choose Your Contribution Scenario

Your situation Recommended scenario
I have a spreadsheet or other tabular data Scenario 1 or Scenario 4
I already have RDF Scenario 2
I have RDF but nowhere to host it Scenario 2 — Route A (we harvest it) or Route B (data portal + federation)
I run my own SPARQL endpoint Scenario 3
I want to add FDOs Scenario 5

Click a scenario to jump directly to the instructions.

📚 Supported Data Types

The MatWerk KG currently supports the following resource types:

🏙️ City
🔬 Materials
🏛️ Organization
👤 People
🗃️ Datasets
💻 Software
🌐 Data Portals
🔭 Instruments
🏗️ Large-scale Facilities
🏷️ Metadata
📐 Ontologies
🎓 Educational Resources
📜 Patents
📦 FDOs
⚙️ Workflows
🔧 Services
🌍 International Collaborations
📅 Events
📄 Publications
🛠️ Tools (NFDI resources)
🔩 OMS Tools
🧪 MatWerk-TA
🔗 MatWerk-IUC
📋 MatWerk-PP
📝 DFG Preface
ℹ️ DFG General Information

🧭 Contribution Scenarios

Scenario 1: Unstructured Data (e.g., spreadsheet not represented with ontology)

You have unstructured data such as a spreadsheet, but you do not want — or do not currently have the time — to model it using an ontology.

This is still a valid way to contribute to the MatWerk KG ecosystem.

⚠️ Please note: We cannot structure the data for you.

What to do

  1. Upload your spreadsheet to Zenodo:
    NFDI MatWerk Community on Zenodo
  2. Provide ROR IDs for organizations and ORCID IDs for people, where applicable.
  3. Your data can then be automatically harvested and added to the MatWerk KG.

⬆ Back to scenario selection

Scenario 2: RDF Data Already Represented with Ontology

You already have RDF data, properly represented using an ontology, and you cannot or do not want to host it yourself. There are two routes, and they differ in one thing: who stores the data.

Route A — we harvest it from where it already lives (GitHub or Zenodo)

Best if your RDF is already published in a repository or a Zenodo record. We fetch the files on a schedule and load each one into its own named graph in the MatWerk KG, so it is queryable alongside everything else and stays linked to the repository it came from.

  1. Publish the RDF where we can reach it:

  2. Register it in the Semantic Dataset Sheet: Semantic Dataset Sheet — one row, and only url is mandatory:

    column what to put needed
    Label a name for the dataset yes
    url the repository URL, Zenodo record URL, or a DOI yes
    source github or zenodo no — inferred from the URL
    files which RDF files to load no — the released artefacts are picked automatically
    Re-sync weekly, monthly or none no
    License, Creator(s) as in the other sheets recommended

  3. Leave files empty unless you need to override the choice. The harvester takes the release assets if the release has any, otherwise the root-level RDF of the repository at the release tag, and collapses OBO-style variants (x.ttl, x-base.ttl, x-full.ttl, x-simple.ttl + the .owl of each) to the full Turtle. Name files explicitly only to pin a different variant, or to load several data files that are not variants of one another.

  4. Open a GitHub issue so we can discuss federated queries, mappings and schema alignment.

After the next run your row is filled in with the graph IRI and a query link. See How harvested RDF is modelled for what is created, and the design patterns for how it is shaped.

Route B — publish it yourself on a data portal, and we federate

Best if you want a citable landing page and your own SPARQL endpoint. The MaterialDigital Data Portal hosts the dataset and gives you a public Fuseki endpoint; we then federate that endpoint rather than copying your data.

  1. Register your dataset at dataportal.material-digital.de (a single RDF file or a whole repository).
  2. Copy the public SPARQL endpoint the portal gives you.
  3. Add it to the SPARQL Endpoint Integration Sheet.
  4. Open a GitHub issue to request federation.

A real one, end to end

Creep Literature Knowledge Graph (CreepLitKG) took Route B — see issue #17:

  • source — GitHub repository
  • landing page — MaterialDigital Data Portal
  • endpoint — https://dataportal.material-digital.de/dataset/a5b4edc4-43ef-44ff-a386-5d1f6fbbc439/fuseki/$/sparql
  • registered in the SPARQL Endpoint Sheet, then federation requested by issue.

Which route?

Route A (we harvest) Route B (portal + federation)
Who stores the data the MatWerk KG, in its own named graph you / the data portal
You need to run a server no no (the portal runs it)
Queryable in the MatWerk KG yes, directly yes, by federation
Stays in step with your source automatically, on a schedule whenever you update the portal
Citable landing page + DOI via Zenodo via the portal

Both are fine. Choose A if the RDF already lives in a repository and you want it inside the KG; choose B if you want your own endpoint and landing page.

⬆ Back to scenario selection

Scenario 3: RDF Data in a Triple Store (Graph Database)

You already maintain your own triple store and provide a SPARQL endpoint. In this case, we can connect your graph to the MatWerk KG so users can query it more seamlessly.

What to do

  1. Add your SPARQL endpoint to the integration sheet:
    SPARQL Endpoint Integration Sheet
  2. Your endpoint can then be automatically integrated into the MatWerk KG.
  3. Optionally, open a GitHub issue to discuss:
  4. mappings
  5. schema alignment
  6. federated SPARQL queries

Submit GitHub Issue

⬆ Back to scenario selection

Scenario 4: Data Type Already Supported in MatWerk KG (e.g., Person, Software)

You want to contribute data about entities that are already supported in the MatWerk KG, such as people, software, or organizations.

What to do

  1. Use the data collection spreadsheet template based on the MatWerk ontology:
    Data Collection Spreadsheet
  2. Fill in your data according to the provided format.
  3. To check whether your data is already available in the knowledge graph, use this spreadsheet:
    Availability Check Spreadsheet

ℹ️ This spreadsheet is connected to a Apache-Airflow workflow and is therefore read-only, which prevents accidental edits from interfering with synchronization.

⬆ Back to scenario selection

Scenario 5: FAIR Digital Objects (FDOs)

You want to add your FAIR Digital Objects (FDOs) to the MatWerk KG.
We provide a simple, semi-automated process so your registered FDOs can be harvested and integrated directly into the graph.

What to do

  1. Add your FDO metadata to the integration spreadsheet:
    FAIR Digital Object Integration Sheet
  2. For each entry, include:
  3. FDO identifier
  4. type
  5. related dataset or publication
  6. persistent URL
  7. Your FDOs will then be automatically harvested and integrated into the MatWerk KG via the FAIR Digital Object Harvester pipeline.
  8. Optionally, open a GitHub issue if you would like to discuss:
  9. custom FDO mappings
  10. schema alignment
  11. FDO–dataset linking strategies

Submit GitHub Issue

⬆ Back to scenario selection

❓ Not Sure Where Your Data Fits?

No problem — we can help.