Seaborn palettes - prevent recycling of colors

2019-03-11 04:08发布

Seaborn allows defining color palettes that contain multiple colors, useful for charts with many lines. However, when setting the palette to one with multiple colors, only the first six are used, after which colors recycle, making it hard to distinguish lines. This can be overridden by explicitly calling the palette, but that's not convenient. Is there a way to force the Seaborn current palette not to recycle colors, when more than 6 are defined?

Example:

from matplotlib import pyplot as plt
import pandas as pd
import seaborn as sb

# Define a palette with 8 colors
cmap = sb.blend_palette(["firebrick", "palegreen"], 8) 
sb.palplot(cmap)

palette with 6 colors

# Set the current palette to this; only 6 colors are used
sb.set_palette(cmap)
sb.palplot(sb.color_palette() )

palette with 6 colors

df = pd.DataFrame({x:[x*10, x*10+5, x*10+10] for x in range(8)})
fig, (ax1, ax2) = plt.subplots(2,1,figsize=(4,6))
# Using the current palette, colors repeat 
df.plot(ax=ax1) 
ax1.legend(bbox_to_anchor=(1.2, 1)) 
# using the palette that defined the current palette, colors don't repeat
df.plot(ax=ax2, color=cmap) 
ax2.legend(bbox_to_anchor=(1.2, 1))  ;

charts with 6 or 8 colors used

1条回答
The star\"
2楼-- · 2019-03-11 04:29

Solution (thanks to @tcaswell for the pointer): Set the palette explicitly using all colors:

# Setting the palette using defaults only finds 6 colors
sb.set_palette(cmap)
sb.palplot(sb.color_palette() )
sb.palplot(sb.color_palette(n_colors=8) )

# but setting the number of colors explicitly allows it to use them all
sb.set_palette(cmap, n_colors=8)
# Even though unless you explicitly request all the colors it only shows 6
sb.palplot(sb.color_palette() )
sb.palplot(sb.color_palette(n_colors=8) )

enter image description here enter image description here enter image description here enter image description here

# In a chart, the palette now has access to all 8 
fig, ax1 = plt.subplots(1,1,figsize=(4,3)) 
df.plot(ax=ax1) 
ax1.legend(bbox_to_anchor=(1.2, 1)) ;

enter image description here

查看更多
登录 后发表回答