Datasets are a fundamental part of the Open Data Cube. A dataset is “The smallest aggregation of data independently described, inventoried, and managed.”​ (Definition of “Granule” from NASA EarthData Unified Metadata Model​)

Examples of Datasets

  • a Landsat Scene​

  • an Albers Equal Area tile portion of a Landsat Scene​

Types of Datasets#

The Open Data Cube supports two primary types of dataset, indexed datasets, and ingested datasets.

Indexed Datasets#

An indexed dataset is available via a file location or from an external uri, with associated metadata available in a format understood by the Data Cube. The pixel data does not need to be stored in the DataCube.


  • USGS Landsat Scenes stored in AWS S3, with prepared agdc-metadata.yaml

  • GA Landsat Scenes

Ingested Datasets#

Data has been created by/and is managed by the Data Cube. The data has typically been been copied, compressed, tiled and possibly re-projected into a shape suitable for analysis, and stored in NetCDF4 files.


  • Tiled GA Landsat Data, ingested into Australian Albers Equal Area Projection (EPSG:3577) and stored in 100km tiles in NetCDF4