Improving Training in Environmental Sciences with Jupyter Notebooks

ECMWF is partner in the EU-funded Earthserver-2 project and explores the feasibility to provide on-demand access to archived meteorological and climate data via the standard interface Web Coverage Service (WCS) defined by the Open Geospatial Consortium (OGC). Providing access to data via OGC web services (e.g. WCS and WMS) has the potential to open up services to to a wider audience, especially to users outside the meteorological and climate community.
Part of the outreach activities of the project was to provide practical hands-on workshops and tutorials to show how users can benefit from OGC web services. Jupyter Notebooks were found to be highly beneficial for geospatial data analysis and for creating reproducible workflows.

Amongst other benefits, Jupyter Notebooks offers the possibility of combining user friendly text, geo data access, manipulation and (interactive) visualisation in one workflow and programming environment. Notebooks can easily be shared via Github to foster team collaboration and reproducible research. JupyterHub, a multi server dedicated to host Jupyter Notebooks, makes trainings very easy, as students can directly start with tackling the actual course problems, rather than spending hours of setting up the right working environment and package installation.

Starting as being a handy outreach tool for the above mentioned project, ECMWF now undertakes efforts to offer more training through Jupyter Notebooks. The presentation will give an overview of the benefits of Jupyter Notebooks and what training ECMWF already provides through Jupyter Notebooks. A special focus will be set on how ECMWF software packages, such as the GRIB decoder ecCodes or the plotting library Magics, can easily be integrated to allow fast manipulation and rendering of meteorological data.