nd – A Framework for the Analysis of n-dimensional Earth Observation Data

Abstract

nd is a software framework for the analysis of n-dimensional datacubes. Such data structures are common in Earth observation where the data are often multivariate on a spatio-temporal grid. The Python library xarray already exists to represent such data sets. This software builds on xarray by providing an interface with other parts of the Python ecosystem, such as scikit-learn and rasterio. Most of the functionality is accessible through an added accessor xarray.Dataset.nd. The software is maintained on GitHub and releases are published on the Python Package Index (PyPI) and Zenodo.

Keywords

• python
• xarray
• earth observation
• remote sensing
• satellite data
• netcdf
• radar
• synthetic aperture radar
• datacube