Plotting Shapely Multipolygon using Matplotlib

Question

I have few polygons or boxes. Ones intersect each other but some are isolated. For example I have free figures:

import shapely.geometry as sg
import shapely.ops as so
import matplotlib.pyplot as plt
r1 = sg.Polygon([(0,0),(0,1),(1,1),(1,0),(0,0)])
r2 = sg.box(0.5,0.5,1.5,1.5)
r3 = sg.box(4,4,5,5)

First two are intersected and r3 is on some distance from them. I unite them via cascaded_union:

new_shape = so.cascaded_union([r1, r2, r3])

Then I try to plot it (one united figure of r1 and r2 and one distanced box r3)

xs, ys = new_shape.exterior.xy
fig, axs = plt.subplots()
axs.fill(xs, ys, alpha=0.5, fc='r', ec='none')
plt.show()

Instead of plot I receive an AttributeError: 'MultiPolygon' object has no attribute 'exterior'. Is there a pythonic way to display a multipolygon or to iterate through it and draw its parts?

Answer 1

Shapely Polygon object has attribute exterior. Shapely MultiPolygon object has Polygon object sequence. You should iterate over those polygons. You can do that using attribute geoms of MultiPolygon.

Use this way:

import shapely.geometry as sg
import shapely.ops as so
import matplotlib.pyplot as plt
r1 = sg.Polygon([(0,0),(0,1),(1,1),(1,0),(0,0)])
r2 = sg.box(0.5,0.5,1.5,1.5)
r3 = sg.box(4,4,5,5)
new_shape = so.unary_union([r1, r2, r3])
fig, axs = plt.subplots()
axs.set_aspect('equal', 'datalim')
for geom in new_shape.geoms:

    xs, ys = geom.exterior.xy

    axs.fill(xs, ys, alpha=0.5, fc='r', ec='none')
plt.show()

Answer 2

An alternative, shorter way of plotting using @Kadir Şahbaz's answer:

Edit: Per @petezurich's comment, cascaded_union is superseded by unary_union as of 1.8.0

new_shape = so.unary_union([r1, r2, r3])
Plot each polygon shape directly
for geom in new_shape.geoms:
    plt.plot(*geom.exterior.xy)
Set (current) axis to be equal before showing plot
plt.gca().axis("equal")
plt.show()

Answer 3

Look at Plot shapefile with islands with matplotlib for example.

As with polygons you can use matplotlib paths and patches and there is a Python module dedicated to plot polygons from shapefiles using these functions Descartes.

new_shape= so.unary_union([r1, r2, r3])
from descartes import PolygonPatch
import matplotlib.pyplot as plt
BLUE = '#6699cc'
GRAY = '#999999'
fig = plt.figure() 
ax = fig.gca() 
ax.add_patch(PolygonPatch(new_shape, fc=GRAY, ec=BLUE, alpha=0.5, zorder=2 ))
ax.axis('scaled')
plt.show()

Answer 4

You can use GeoPandas' .plot() function:

import shapely.geometry as sg
import shapely.ops as so
import matplotlib.pyplot as plt
import geopandas as gpd
r1 = sg.Polygon([(0, 0), (0, 1), (1, 1), (1, 0), (0, 0)])
r2 = sg.box(0.5, 0.5, 1.5, 1.5)
r3 = sg.box(4, 4, 5, 5)
new_shape = so.cascaded_union([r1, r2, r3])
gpd.GeoSeries(new_shape).plot()
plt.show()