Administrative boundaries — Natural Earth & US Census TIGER¶

Beyond per-country geoBoundaries, the admin backend serves global cultural layers (Natural Earth) and US Census TIGER/Line cartographic boundaries. This notebook maps both and shows the automatic CRS normalisation to EPSG:4326.

Setup¶

In [ ]:

import matplotlib.pyplot as plt

from earthlens.core import EarthLens

import matplotlib.pyplot as plt from earthlens.core import EarthLens

Natural Earth — world countries¶

Natural Earth ships cultural layers at three scales (10m / 50m / 110m). The scale= selector is optional and defaults per dataset (110m for countries). We use the coarse 110m layer here — it is tiny (~200 KB) and fast.

In [ ]:

countries = EarthLens(
    data_source='natural-earth',
    variables=['natural_earth:countries'],
    scale='110m',
).download(progress_bar=False)
print(len(countries), 'countries, EPSG:', countries.crs.to_epsg())

countries = EarthLens( data_source='natural-earth', variables=['natural_earth:countries'], scale='110m', ).download(progress_bar=False) print(len(countries), 'countries, EPSG:', countries.crs.to_epsg())

In [ ]:

ax = countries.plot(edgecolor='white', facecolor='#9ecae1', figsize=(11, 5.5))
ax.set_title('Natural Earth — admin-0 countries (110m)')
ax.set_xlabel('Longitude')
ax.set_ylabel('Latitude')
plt.show()

ax = countries.plot(edgecolor='white', facecolor='#9ecae1', figsize=(11, 5.5)) ax.set_title('Natural Earth — admin-0 countries (110m)') ax.set_xlabel('Longitude') ax.set_ylabel('Latitude') plt.show()

US Census TIGER — states¶

TIGER cartographic-boundary files arrive in NAD83 (EPSG:4269); the backend reprojects them to EPSG:4326 so every admin result shares one CRS. The year= selector is optional (defaults to the dataset's vintage).

In [ ]:

states = EarthLens(
    data_source='tiger',
    variables=['tiger:state'],
).download(progress_bar=False)
print(len(states), 'state features, EPSG:', states.crs.to_epsg())

states = EarthLens( data_source='tiger', variables=['tiger:state'], ).download(progress_bar=False) print(len(states), 'state features, EPSG:', states.crs.to_epsg())

The 56 features include the 50 states plus DC and territories. We clip the plot to the contiguous US for readability.

In [ ]:

ax = states.plot(edgecolor='black', facecolor='#fdd0a2', figsize=(10, 6))
ax.set_xlim(-128, -65)
ax.set_ylim(23, 50)
ax.set_title('US Census TIGER — states (reprojected NAD83 -> EPSG:4326)')
ax.set_xlabel('Longitude')
ax.set_ylabel('Latitude')
plt.show()

ax = states.plot(edgecolor='black', facecolor='#fdd0a2', figsize=(10, 6)) ax.set_xlim(-128, -65) ax.set_ylim(23, 50) ax.set_title('US Census TIGER — states (reprojected NAD83 -> EPSG:4326)') ax.set_xlabel('Longitude') ax.set_ylabel('Latitude') plt.show()

A note on CGAZ¶

The fourth source, CGAZ (Comprehensive Global Admin Zones), provides seamless global ADM0/1/2 layers — every country in one file. Those files are large (160–550 MB), so they are not fetched in this tutorial; use variables=['cgaz:adm0'] (no selector) when you need the seamless global layer. It returns polygons in EPSG:4326 like the others.

Takeaway¶

• Natural Earth (scale=) and TIGER (year= / state=) follow the same download() shape as geoBoundaries.
• The backend normalises every source to EPSG:4326 — TIGER's NAD83 is reprojected automatically.
• aggregate= is rejected for all admin datasets: boundaries are vector polygons, not a gridded field.