import palettable
bmap = palettable.tableau.Tableau_20.mpl_colors
all_markers = ['s', 'o', '^', '*', 'v'] * 100
max_values = len(bmap) if len(bmap) < len(all_markers) else len(all_markers)
# Make sure equal number of cycle elements for color and markers
markers = all_markers[:max_values]
bmap = bmap[:max_values]
color_cycle = cycler('color', bmap) # color cycle
marker_cycle = cycler('marker', markers) # marker cycle
plt.rc('axes', prop_cycle=(color_cycle + marker_cycle))
I use the above code to cycle through a combination of colors and markers for some plots. However, this messes up a seaborn swarmplot I am using by replacing the dot markers. The dataframe used in boxplot and swarmplot is this:
Poor Watch Favourable Exceptional
0 0.256667 0.186000 0.100667 0.456667
1 0.259333 0.150000 0.181333 0.409333
2 0.360000 0.200667 0.158667 0.280667
3 0.380667 0.217333 0.109333 0.292667
4 0.258667 0.141333 0.150667 0.449333
5 0.210000 0.146000 0.167333 0.476667
6 0.794667 0.052000 0.056000 0.097333
7 0.269333 0.157333 0.112000 0.461333
8 0.338667 0.230667 0.119333 0.311333
9 0.716000 0.092000 0.084000 0.108000
10 0.627333 0.103333 0.124000 0.145333
11 0.332667 0.171333 0.098667 0.397333
12 0.331333 0.210000 0.122667 0.336000
13 0.417333 0.164667 0.108000 0.310000
The code is here:
# Copy the data above
import pandas
import matplotlib.pyplt as plt
import seaborn as sns
df = pandas.read_clipboard()
ax = sns.boxplot(data=df, linewidth=1.5, width=0.25)
# stackoverflow.com/questions/34163622/seaborn-passes-kwargs-to-plt-boxplot
plt.setp(ax.artists, alpha=.3, fill=False)
ax = sns.swarmplot(data=df, color='.25')
plt.setp(ax.artists, alpha=.6)
ax.set_ylabel('Class probability')
ax.set(ylim=(0, 1.0))
plt.tight_layout()
plt.show()
Here is the resulting graphic. I want to ensure that the dot markers are not replaced. How do I do that?
