Calculate country level aggregations by combining gridded data with geometry data - ECMWF internal earthkit training, March 2026

Using only the packages imported and data loaded in the first code block, produce a figure similar to the one show below.

All the information and code required to complete this task can be found in the earthkit-transforms Documentation pages.

The following code cell imports the required earthkit modules and uses earthkit-data to download the sample data. There are two data objects:

era5_data is a grib file containing hourly 2m temperature data over Europe for the 1st January 2015.
nuts_data is a geojson containing the country borders accroding to the Nomenclature of territorial units for statistics.

from earthkit.data.utils.testing import earthkit_remote_test_data_file

from earthkit import data as ekd
from earthkit import plots as ekp
from earthkit import transforms as ekt

# Get some demonstration ERA5 data, this could be any url or path to an ERA5 grib or netCDF file.
# remote_era5_file = earthkit_remote_test_data_file("era5_temperature_europe_2015.grib") # Large file
remote_era5_file = earthkit_remote_test_data_file("era5_temperature_europe_20150101.grib")
era5_data = ekd.from_source("url", remote_era5_file)

# Use some demonstration polygons stored, this could be any url or path to geojson file
remote_nuts_url = earthkit_remote_test_data_file("NUTS_RG_60M_2021_4326_LEVL_0.geojson")
nuts_data = ekd.from_source("url", remote_nuts_url).to_geopandas()

# Insert your code here

Hints¶

Consider the following workflow:

Optional, convert to xarray to have more control on the dimensionality of the results.
Use the earthkit.transforms.spatial sub-module to calculate the country averages.
- The time dimension should be averaged over at the same time as the spatial dimensions
- Ensure the statistic you calculate is weighted by latitude
- Return the result as a pandas DataFrame for simpler plotting
Use earthkit.plots.geo.choropleth to create the plot