CODEWITHSUNDEEP
pythonpandasmatplotlibgeopandas
Lucas H
Lucas H

Reputation: 1027

Custom Choropleth Labels in Geopandas

I am attempting to make a choropleth plot in geopandas. Because of the large number of objects and a highly skewed distribution, I am binning into quantiles using the built in mapclassify support. It is working great, except the labels for the resulting legend are terrible.

As a test case, I have produced the following:

enter image description here

Using the following code:

import geopandas as gp
import matplotlib.pyplot as plt

world = gp.read_file(gp.datasets.get_path('naturalearth_lowres'))
fig, ax = plt.subplots()
world.plot(ax=ax,column='pop_est',
           cmap='OrRd',
           scheme='quantiles',
           legend=True,
           legend_kwds=dict(loc='upper right',
                            bbox_to_anchor=(1.5, .9),
                            fontsize='small',
                            title="Legend",
                            frameon=False)
           )
ax.axis('off')

I want at a minimum to be able to have fully custom labels for the legend, for example:

However, the ideal case would be to have a continuous blocked colorbar with custom labels, for example, like the following US map (which I did not produce):

enter image description here

Upvotes: 2

Views: 1719

Answers (1)

Mattijn
Mattijn

Reputation: 13880

It was not directly obvious from the comments how to do this. The following, bit verbose piece of code produce something with a custom legend:

import geopandas as gp
import matplotlib.pyplot as plt

def update_custom_legend_labels(ax, new_labels):
    current_legend = ax.get_legend()
    for ix_current_label, txt_current_label in enumerate(current_legend.texts):    
        for ix_new_label, txt_new_label in new_labels.items():        
            if ix_current_label == ix_new_label:
                # print(f'current: {ix_current_label, txt_current_label}')
                # print(f'new: {ix_new_label, txt_new_label}')
                txt_current_label.set_text(txt_new_label)
                # print(f'update: {txt_current_label}')

world = gp.read_file(gp.datasets.get_path('naturalearth_lowres'))
fig, ax = plt.subplots()
world.plot(ax=ax,
           column='pop_est',
           cmap='OrRd',
           scheme='quantiles',
           legend=True,
           legend_kwds=dict(loc='upper right',
                            bbox_to_anchor=(1.5, .9),
                            fontsize='small',
                            title="Legend",
                            frameon=False, fmt="{:.0f}")
           )

current_labels = [t.get_text() for t in ax.get_legend().get_texts()]

no_labels = len(current_labels)
new_labels = []
for ix, val in enumerate(current_labels):    
    if ix == 0:  # first item
        val = '<' + val.split(',')[0]
    elif no_labels == ix + 1:  # last item
        val = '>' + val.split(',')[0]        
    else:  # remaining items
        val = val.replace(',', ' -')            
    new_labels.append((ix,val))
new_labels = dict(new_labels)    

update_custom_legend_labels(ax, new_labels)
ax.set_axis_off()
plt.show()

enter image description here

Upvotes: 1

Related Questions