Which data to use for geopandas.GeoDataFrame choropleth example?

[michaelgrund]

I wanted to continue on the choropleth example mentioned in #1374, however, in the meanwhile
world = gpd.read_file(gpd.datasets.get_path("naturalearth_lowres")) raises a deprection warning saying the datasets will be removed in geopandas v1.0 (see discussion in #2751).

Thus, we need to use another data source. Any ideas what can be used without to much effort and without adding further dependecies to PyGMT?

[seisman]

Maybe read a URL directly?

world = gpd.read_file("https://naciscdn.org/naturalearth/110m/physical/ne_110m_land.zip")

[michaelgrund]

Maybe read a URL directly?

world = gpd.read_file("https://naciscdn.org/naturalearth/110m/physical/ne_110m_land.zip")

Yes that works but unfortunately the fetched data does neither contain population values nor it's possible to filter e.g. for a specific continent/region like "Europe".

[yvonnefroehlich]

I faced this issue when starting to work on possible figures for the PyGMT paper.

So far I understood that there is now a separate repro from the geopandas organization for datasets: geodatasets https://github.com/geopandas/geodatasets and https://geodatasets.readthedocs.io/en/latest/.

These datasets can be accessed:

import geopandas
import geodatasets

airbnb = geopandas.read_file(geodatasets.get_path("geoda.airbnb"))

It is possible to get more information about these datasets (please see https://geodatasets.readthedocs.io/en/latest/#how-to-use):

import geodatasets
geodatasets.data

Gives:

{'geoda': {'airbnb': {'url': 'https://geodacenter.github.io/data-and-lab//data/airbnb.zip',
   'license': 'NA',
   'attribution': 'Center for Spatial Data Science, University of Chicago',
   'name': 'geoda.airbnb',
   'description': 'Airbnb rentals, socioeconomics, and crime in Chicago',
   'geometry_type': 'Polygon',
   'nrows': 77,
   'ncols': 21,
   'details': 'https://geodacenter.github.io/data-and-lab//airbnb/',
   'hash': 'a2ab1e3f938226d287dd76cde18c00e2d3a260640dd826da7131827d9e76c824',
   'filename': 'airbnb.zip'},
  'atlanta': {'url': 'https://geodacenter.github.io/data-and-lab//data/atlanta_hom.zip',
   'license': 'NA',
   'attribution': 'Center for Spatial Data Science, University of Chicago',
   'name': 'geoda.atlanta',
   'description': 'Atlanta, GA region homicide counts and rates',
   'geometry_type': 'Polygon',
   'nrows': 90,
   'ncols': 24,
   'details': 'https://geodacenter.github.io/data-and-lab//atlanta_old/',
   'hash': 'a33a76e12168fe84361e60c88a9df4856730487305846c559715c89b1a2b5e09',
   'filename': 'atlanta_hom.zip',
   'members': ['atlanta_hom/atl_hom.geojson']},
   ...

There also the complete URL can be found (directly via geodatasets.data.geoda.airbnb or geodatasets.get_url("geoda.airbnb")). The URL can be passed directly to geopandas.read_file(). So, if you know the URL of the dataset, it is not needed anymore to have geodatasets installed.

The dataset is read into a GeoDataFrame and can then be used for analysis and plotting, e.g., in PyGMT. I just tried this with the first dataset listed by geodatasets.data:

import geopandas as gpd
import numpy as np
import pygmt

airbnb = gpd.read_file("https://geodacenter.github.io/data-and-lab//data/airbnb.zip")

fig = pygmt.Figure()

fig.coast(
    region=[-88, -87.2, 41.6, 42.05],
    projection="M10c", 
    frame=True,
    water="lightblue",
    land="gray90",
    shorelines="1/1p,gray30",
)

pygmt.makecpt(
    cmap="bilbao",
    series=[np.min(airbnb["population"]), np.max(airbnb["population"]), 10],
    continuous=True,
)

fig.plot(
    data=airbnb[["population","geometry"]], 
    pen="0.2p,gray10", 
    close=True, 
    fill="+z", 
    cmap=True,
    aspatial="Z=population",
)

fig.colorbar(frame="x+lPopulation")

fig.show()
# fig.savefig(fname="choropleth_airbnb_population.png")

Output figure:

So, @michaelgrund maybe you can find in the overview at https://geodatasets.readthedocs.io/en/latest/introduction.html#what-is-the-geodatasets-data-object an interesting dataset which you can use for your gallery example?

[michaelgrund]

I didn't realize that I could load the file directly. I thought you also need to include the geodatasets library, which would have been a new dependency. Thanks @yvonnefroehlich. Just modified your example a little, see #2796.