Closed
Description
In the example scatter_piecharts.py, segments of the unit disk are approximated through polygons and used as markers. By chance, each segment has a coordinate with modulus 1, so that it is not rescaled. As soon as one starts to play with the example a bit, e.g. changing the proportions of the first two segments into
r1 = 0.1
r2 = r1 + 0.05
the second, green segment is rescaled, which it should not be for our purposes.
To counter that effect, one needs to apply a segment-specific scale, as is done here:
The resulting code looks like this (there's probably a nicer waying of writing this):
"""
This example makes custom 'pie charts' as the markers for a scatter plotqu
Thanks to Manuel Metz for the example
"""
import math
import numpy as np
import matplotlib.pyplot as plt
# first define the ratios
r1 = 0.2 # 20%
r2 = r1 + 0.4 # 40%
# define some sizes of the scatter marker
sizes = [60,80,120]
# calculate the points of the first pie marker
#
# these are just the origin (0,0) +
# some points on a circle cos,sin
x = [0] + np.cos(np.linspace(0, 2*math.pi*r1, 10)).tolist()
y = [0] + np.sin(np.linspace(0, 2*math.pi*r1, 10)).tolist()
xy1 = list(zip(x,y))
s1 = max(max(x),max(y))
# ...
x = [0] + np.cos(np.linspace(2*math.pi*r1, 2*math.pi*r2, 10)).tolist()
y = [0] + np.sin(np.linspace(2*math.pi*r1, 2*math.pi*r2, 10)).tolist()
xy2 = list(zip(x,y))
s2 = max(max(x),max(y))
x = [0] + np.cos(np.linspace(2*math.pi*r2, 2*math.pi, 10)).tolist()
y = [0] + np.sin(np.linspace(2*math.pi*r2, 2*math.pi, 10)).tolist()
xy3 = list(zip(x,y))
s3 = max(max(x),max(y))
fig, ax = plt.subplots()
ax.scatter( np.arange(3), np.arange(3), marker=(xy1,0),
s=[ s1*s1*x for x in sizes ], facecolor='blue' )
ax.scatter( np.arange(3), np.arange(3), marker=(xy2,0),
s=[ s2*s2*x for x in sizes ], facecolor='green' )
ax.scatter( np.arange(3), np.arange(3), marker=(xy3,0),
s=[ s3*s3*x for x in sizes ], facecolor='red' )
plt.show()Metadata
Metadata
Assignees
Labels
No labels