How can I add a legend while plotting multiple geopandas dataframes in the same subplot?

Question

I have a geopandas dataframe world which I created using:

import geopandas as gpd

world = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))

I created two different geodataframes for usa and china as shown below:

usa = world[world.name == "United States of America"]

china = world[world.name == "China"]

I want to plot the USA as blue and China as red in the map. I plotted it using the following line of code:

fig, ax = plt.subplots(figsize = (20, 8))
world.plot(ax = ax, color = "whitesmoke", ec = "black")
usa.plot(ax = ax, color = "blue", label = "USA")
china.plot(ax = ax, color = "red", label = "China")
ax.legend()
plt.show()

It looks as follows:

I want to add legends stating blue for the USA and red for China. Therefore, I gave labels as shown in the code above. However, I get the following warning:

No artists with labels found to put in legend. Note that artists whose label start with an underscore are ignored when legend() is called with no argument.

I am not able to add the legend. How can I add the legends for the USA and China in this plot? Is it possible using geopandas and matplotlib?

Answer 1

I never used geopandas, however looking at the result is appears that those filled areas are PathCollection, which are not supported on legends. But we can create legend artists:

import geopandas as gpd
from matplotlib.lines import Line2D

world = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))
usa = world[world.name == "United States of America"]
china = world[world.name == "China"]

fig, ax = plt.subplots()
world.plot(ax = ax, color = "whitesmoke", ec = "black")
usa.plot(ax = ax, color = "blue", label = "USA")
china.plot(ax = ax, color = "red", label = "China")

lines = [
    Line2D([0], [0], linestyle="none", marker="s", markersize=10, markerfacecolor=t.get_facecolor())
    for t in ax.collections[1:]
]
labels = [t.get_label() for t in ax.collections[1:]]
ax.legend(lines, labels)
plt.show()