DBCells Vocabulary

A lightweight RDF/RDFS vocabulary for publishing, reusing and integrating data from land use and land cover change models as linked data.

This vocabulary is part of the DBCells project, which proposes a methodology for reproducible spatial dynamic models based on the linked data paradigm.

---

Canonical URI

https://purl.org/linked-data/dbcells

⚠️ The code examples in the paper published at V Simpósio REACT (2025) contain incorrect URIs using http://www.purl.org/.... The correct base is always https://purl.org/... (no www, HTTPS).

---

Modules

The vocabulary is split into four modules, each with its own PURL and documentation page:

Module	PURL	Documentation	Description
Root	/dbcells	HTML	Cell class and core properties
Measures	/dbcells/measure	HTML	Spatial aggregation operations
Attributes	/dbcells/attribute	HTML	Metadata qualifying observations
Code lists	/dbcells/code	HTML	Geographic features and land cover classes

---

Design

The vocabulary is built on top of the RDF Data Cube Vocabulary (W3C recommendation) and organises data into three components:

• Dimensions — identify each observation in space and time, reusing sdmx:refArea and sdmx:refPeriod from the SDMX vocabulary
• Measures — represent the spatial aggregation operation applied to the data (e.g. mean, distance, percentage)
• Attributes — qualify each observation by linking it to the geographic feature, source file and script used to compute it

The key design decision is to separate the measure from what is being measured. For example, "mean of forest cover" is modelled as:

• measure: dbc-m:mean
• attribute: dbc-a:feature → dbc-c:landcover-veg

This allows measures to be reused across different models while attributes carry the domain-specific meaning.

---

Quick start

@prefix qb:       <http://purl.org/linked-data/cube#> .
@prefix sdmx-dim: <http://purl.org/linked-data/sdmx/2009/dimension#> .
@prefix dcterms:  <http://purl.org/dc/terms/> .
@prefix xsd:      <http://www.w3.org/2001/XMLSchema#> .
@prefix dbc-m:    <https://purl.org/linked-data/dbcells/measure#> .
@prefix dbc-a:    <https://purl.org/linked-data/dbcells/attribute#> .
@prefix dbc-c:    <https://purl.org/linked-data/dbcells/code#> .
@prefix cells:    <https://purl.org/linked-data/dbcells/epsg4326#> .
@prefix ds:       <https://purl.org/dbcells/dataset#> .
@prefix obs:      <https://purl.org/dbcells/observation#> .

# Dataset — groups common metadata for a set of observations
ds:forest-cover-2010
    a                   qb:DataSet ;
    dcterms:title       "Forest vegetation cover — Brazil 2010"@en ;
    sdmx-dim:refPeriod  "2010"^^xsd:gYear ;
    dbc-a:feature       dbc-c:landcover-veg ;
    dbc-a:sourceFile    <https://github.com/LambdaGeo/brlucc-database/raw/main/data/mapbiomas_2010.tif> ;
    dbc-a:scriptFile    <https://github.com/LambdaGeo/brlucc-database/blob/main/scripts/fill_forest_percentage.lua> .

# Observation
obs:forest-cover-001
    a                   qb:Observation ;
    qb:dataSet          ds:forest-cover-2010 ;
    sdmx-dim:refArea    cells:R0_0830Cx-46_4583Cy-23_8881 ;
    dbc-m:percentage    87.3 .

See examples/example.ttl for a complete example with multiple datasets, observations and a SPARQL query.

---

Repository structure

.
├── vocab/
│   ├── dbcells.ttl          # root — Cell class and core properties
│   ├── dbc-measure.ttl      # measure properties
│   ├── dbc-attribute.ttl    # attribute properties
│   └── dbc-code.ttl         # code lists
├── examples/
│   └── example.ttl          # usage example
├── scripts/
│   ├── gendoc.py            # generates docs/ using pyLODE
│   ├── Pipfile
│   └── Pipfile.lock
├── docs/                    # auto-generated — do not edit manually
└── archive/                 # legacy files — do not use

---

Generating documentation

Documentation is generated automatically via pyLODE on every push to main using GitHub Actions. The generated HTML pages are published via GitHub Pages at:

https://lambdageo.github.io/dbcells-ontology/

To regenerate locally:

cd scripts/
pipenv install
pipenv run python gendoc.py

---

Published data

Data from land use and land cover models published using this vocabulary is available at:

Dataset	URL
Land cover — Brazil	https://data.world/lambdageo/luccmebrlanduse
Drivers (distance features)	https://data.world/lambdageo/luccmebrdrivers
DBCells cell repository	https://data.world/dbcells/dbcells

---

Related projects

• DBCells server — linked data server for cell geometries
• brlucc-database — raw data and scripts for the Brazil LUCC model
• QGISSPARQL — QGIS plugins for importing and exporting linked data

---

License

This vocabulary is published under the Creative Commons Attribution 4.0 International License.

---

Citation

If you use this vocabulary in your work, please cite:

Santos Junior, N. J.; Silva, C. D. S.; Sousa, F. M.; Bezerra, D. S.; Costa, S. S. Publishing and Reusing Land Use and Land Cover Data through RDF Data Cube. V Simpósio REACT, 2025.