GHSL — building height¶

Download GHS-BUILT-H (average net building height, ANBH) for 2018 at 100 m over a Lisbon AOI, then render the height raster with matplotlib. GHSL is a raster backend, so download() writes a GeoTIFF to disk and returns its path.

Setup¶

Imports up front: tempfile for a scratch download directory, numpy / matplotlib for the plot, pyramids' Dataset for reading the GeoTIFF, and earthlens for the unified EarthLens entry point.

In [ ]:

import tempfile

import matplotlib.pyplot as plt
import numpy as np
from pyramids.dataset import Dataset

from earthlens import EarthLens

import tempfile import matplotlib.pyplot as plt import numpy as np from pyramids.dataset import Dataset from earthlens import EarthLens

Download the building-height tile¶

We build the GHSL request first — source, the built_height variable, the 2018 year window, and a bounding box over Lisbon. The output goes to a temporary directory created with tempfile.mkdtemp().

In [ ]:

out = tempfile.mkdtemp()
ghsl = EarthLens(
    data_source='ghsl',
    variables=['built_height'],
    start='2018-01-01',
    end='2018-12-31',
    aoi=[-9.20, 38.68, -9.10, 38.78],
    path=out,
)

out = tempfile.mkdtemp() ghsl = EarthLens( data_source='ghsl', variables=['built_height'], start='2018-01-01', end='2018-12-31', aoi=[-9.20, 38.68, -9.10, 38.78], path=out, )

Downloading is kept as its own step: download() fetches the tile, crops it to the AOI, and returns the list of written GeoTIFF paths.

In [ ]:

paths = ghsl.download(progress_bar=False)
paths

paths = ghsl.download(progress_bar=False) paths

Read and clean the height array¶

Read the GeoTIFF into a NumPy array with pyramids' Dataset, squeeze the single band, and mask the no-data sentinel (negative values) to NaN so they render as blank rather than skewing the colour scale.

In [ ]:

raster = Dataset.read_file(str(paths[0]))
arr = np.asarray(raster.read_array(), dtype='float32')
arr = arr[0] if arr.ndim == 3 else arr
arr = np.where(arr < 0, np.nan, arr)

raster = Dataset.read_file(str(paths[0])) arr = np.asarray(raster.read_array(), dtype='float32') arr = arr[0] if arr.ndim == 3 else arr arr = np.where(arr < 0, np.nan, arr)

Plot the building heights¶

Render the average net building height over Lisbon with a viridis colour map.

In [ ]:

plt.figure(figsize=(6, 5))
plt.imshow(arr, cmap='viridis')
plt.colorbar(label='mean height (m)')
plt.title('GHS-BUILT-H ANBH 2018 (100 m) — Lisbon')
plt.axis('off')
plt.show()

plt.figure(figsize=(6, 5)) plt.imshow(arr, cmap='viridis') plt.colorbar(label='mean height (m)') plt.title('GHS-BUILT-H ANBH 2018 (100 m) — Lisbon') plt.axis('off') plt.show()