CODEWITHSUNDEEP
pythonpython-3.xmatplotlibgeospatialgeopandas
hbstha123
hbstha123

Reputation: 1636

How can I get legend when I specify user-defined color for plotting maps using geopandas?

I have created the map of world from inherent datasets of geopandas as follows:

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

I have plotted the map of world by highlighting the unique continents using

world.plot("continent", legend = True)

The legends are also present in the plot. It looks as follows: enter image description here

I get unique color for each continent automatically. However, I'd like to define the color for each continent myself as well as have them in the legends. To define unique color for each continent, I created a dictionary color_dict which looks as follows:

{'Oceania': 'yellow',
 'Africa': 'brown',
 'North America': 'maroon',
 'Asia': 'blue',
 'South America': 'green',
 'Europe': 'red',
 'Seven seas (open ocean)': 'skyblue',
 'Antarctica': 'white'}

And then I appended it as a new column to the world geopandas dataframe and created the plot again.

world["Colors"] = world["continent"].map(color_dict)
world.plot("continent",
           color = world["Colors"],
           legend = True)

enter image description here

I get the desired color for each continent. However, now I am not getting the continents name and color in the legend. How can I bring the legends in the plot when I define the color myself?

Upvotes: 0

Views: 1208

Answers (2)

hbstha123
hbstha123

Reputation: 1636

Based on this, this is the potential solution:

color_dict = {'Oceania': 'yellow',
 'Africa': 'brown',
 'North America': 'maroon',
 'Asia': 'blue',
 'South America': 'green',
 'Europe': 'red',
 'Seven seas (open ocean)': 'skyblue',
 'Antarctica': 'white'}

world["Colors"] = world["continent"].map(color_dict)

ax = world.plot(color = world["Colors"])

# add manual legend
from matplotlib.lines import Line2D
custom_points = [Line2D([0], [0], marker="o", linestyle="none", markersize=10, color=color) for color in color_dict.values()]
leg_points = ax.legend(custom_points, color_dict.keys())
ax.add_artist(leg_points)

Output: enter image description here

Upvotes: 0

ekrall
ekrall

Reputation: 192

You may find that you have better results, and greater control, if you convert the geopandas geometries to polygon patches using descartes, and then generate the plots using matplotlib.

Plotting from geopandas can be finicky. it may take some trial and error to get exactly what you want.

Ensure that you have these packages in your module

from matplotlib import pyplot
from descartes import PolygonPatch
from geopandas import GeoDataFrame

PolygonPatch allows you to specify an edge color and face color for each geometry.

With pyplot you should be able to create the legend you desire.

Form your geodataframe (gdf) of the continents. For each continent in your gdf, form a PolygonPatch with the desired face color. Then add the patches to a pyplot figure. With pyplot, you should be able to specify a legend and manipulate it more easily.

Here is some documentation on PolygonPatches, and an example of plotting them (albeit not with the exact legend formatting you are seeking): https://pypi.org/project/descartes/ https://geoffboeing.com/2014/09/visualizing-summer-travels-part-6-projecting-spatial-data-python/

Upvotes: 1

Related Questions