Lexcube for Jupyter: Interactive Earth System Data Cube Visualization in Jupyter Notebooks

Data streams representing the Earth system both through modeling and remote sensing approaches, encompass a diverse range and massive amount of information. Unveiling insights at global and local scales becomes increasingly challenging for the wider public and the broader scientific audience as the temporal and spatial resolutions of data sets continually improve. An effective solution to this involves the development of fully interactive visualizations capable of rendering terabytes of data in real-time, spanning time, space, variables, and model variants. Lexcube.org, the Leipzig Explorer of Earth Data Cubes, was the first tool that allowed to explore and interact with large Earth system data sets in the form of an interactive data cube visualization in the web browser, but was limited to a few preset data sets.

Here we present Lexcube for Jupyter, a Jupyter notebook extension building on top of the existing Lexcube.org software components, that allows to visualize any spatiotemporal or otherwise three-dimensional data as an interactive 3D data cube. The data cube visualization treats all three dimensions equally and, e.g., in the case of a spatiotemporal data cube, allows to inspect temporal patterns in a novel way. Interaction with the data cube is designed to be intuitive, also allowing touch gestures on touch-capable devices. Building on top of the powerful open-source libraries Xarray and Numpy, Lexcube for Juypter integrates effortlessly into the existing ecosystem of open-source data cube software components as it is able to visualize any gridded data set from those libraries, including remotely stored and chunked data sets. Furthermore, Lexcube for Jupyter allows to export the currently visible data cube as a new Xarray or Numpy object, allowing scientists to use Lexcube in their workflow for data selection and curation. In addition, new disciplines such as the atmospheric sciences may profit from Lexcube for Juypter as they can now visualize their own three-dimensional data that is not necessarily spatiotemporal, e.g., three-dimensional atmospheric humidity data cubes (latitude×longitude×pressure level) as seen on lexcube.org. Lexcube for Jupyter is open-source and available on GitHub and PyPi since January 2024.