Day 4

Detailed paper information

Back to list

Paper title Digital Earth Australia’s Notebooks and Tools repository: an open-source repository showcasing reproducible Earth observation data science
Authors
  1. Claire Krause Geoscience Australia Speaker
  2. Bex Dunn Geoscience Australia
  3. Robbi Bishop-Taylor Geoscience Australia
  4. Caitlin Adams FrontierSI
  5. Chad Burton Geoscience Australia
  6. Matthew Alger Geoscience Australia
  7. Sean Chua Geoscience Australia
  8. Claire Phillips Geoscience Australia
  9. Vanessa Newey Geoscience Australia
  10. Kirill Kouzoubov Geoscience Australia
  11. Alex Leith Geoscience Australia
  12. Damien Ayers Geoscience Australia
  13. Andrew Hicks FrontierSI
Form of presentation Poster
Topics
  • Open Earth Forum
    • C5.03 Open Source, data science and toolboxes in EO: Current status & evolution
Abstract text Digital Earth Australia (DEA) is a government program that enables government, industry, and academia to more easily make use of Earth observation data in Australia. DEA does this by producing and disseminating free and open analysis-ready data from the Landsat and Sentinel-2 missions. Data are processed, stored and indexed into an instance of the Open Data Cube, enabling API-based access to more than thirty years of Earth observation (EO) data.

Making EO data simply available is not enough. Users need to be able to investigate specific applications. Barriers to applying satellite imagery include uncertainty in how the data can be applied to the application, difficulties in accessing the data, and challenges in analysing the petabytes of available data. The DEA notebooks and tools repository on GitHub ("DEA Notebooks") hosts Jupyter notebooks, Python scripts and workflows for analysing DEA satellite data and its derived products. The repository provides a guide to getting started with DEA and showcases the wide range of geospatial analyses that can be achieved using open-source software including the Open Data Cube and xarray. The DEA Notebooks repository steps users through an introduction to the Python packages needed to analyse data and introduces datasets available through DEA. It provides frequently used code snippets for quick tasks such as creating animations or masking data as well as more complex workflows such as machine learning and production of derived products.

A community of practice has evolved around the DEA Notebooks repository. The repository is regularly maintained and updated and meets clearly defined standards of quality, enabled by templates for contributors. A Wiki and user guide are also provided to assist users with accessing DEA, as well as channels for seeking support. Workflows are built upon and committed back to the repository for other users to benefit from. DEA Notebooks has been utilised to teach multiple University degree-level courses across Australia, underpinned peer-reviewed scientific publications, and facilitated two digital art projects. The DEA Notebooks project evolved to drive documentation and user engagement for the DEA program as a whole and is now a rich resource for new and existing users of Earth observation datasets.

The repository can be accessed at https://github.com/GeoscienceAustralia/dea-notebooks.