First steps to create a datacube

To create a datacube using Earth Data Store from EarthDaily you have two possibilities.

The first one is the more classic one, you request items, then you build your datacube, and then you can mask pixels using a cloudmask asset. The second one is the most turnkey one, the one we recommend, it allows to do all the process at once.

Import librairies

from matplotlib import pyplot as plt

import earthdaily
from earthdaily import EarthDataStore, datasets

Loading geometry

geometry = datasets.load_pivot()

Init earthdatastore with environment variables or default credentials

eds = EarthDataStore()

Create datacube (in one step)

s2_datacube = eds.datacube(
    "sentinel-2-l2a",
    assets=["blue", "green", "red", "nir"],
    intersects=geometry,
    datetime=["2022-08-01", "2022-08-09"],
    mask_with="native",  # equal to "scl" for sentinel-2
    clear_cover=50,
)

s2_datacube.clear_percent.plot.scatter(x="time")
plt.title("Percentage of clear pixels on the study site")
plt.show()

s2_datacube.ed.plot_rgb(vmin=0, vmax=0.2, col="time", col_wrap=4)

• 
• 

<xarray.plot.facetgrid.FacetGrid object at 0x7f5fd00dc350>

Create datacube in three steps

Request items

items = eds.search(
    "sentinel-2-l2a",
    intersects=geometry,
    datetime=["2022-08-01", "2022-08-09"],
)

Creata datacube (independent from being log into earthdatastore) We request the “scl” assets which is the native cloudmask

s2_datacube = earthdaily.earthdatastore.datacube(
    items, assets=["blue", "green", "red", "nir", "scl"], intersects=geometry
)

Mask datacube

# intersects or bbox are asked in order to compute accurate mask statistics

s2_datacube = earthdaily.earthdatastore.mask.Mask(s2_datacube, intersects=geometry).scl(
    mask_statistics=True
)

s2_datacube = earthdaily.earthdatastore.mask.filter_clear_cover(
    s2_datacube, 50
)  # at least 50% of clear pixels
#
s2_datacube.ed.plot_rgb(vmin=0, vmax=0.2, col="time", col_wrap=4)

<xarray.plot.facetgrid.FacetGrid object at 0x7f5fc4392850>