python-3.x - Plot crosstab results using All row as benchmark lines

Question

I have this sample dataframe:

test = pd.DataFrame({'cluster':['1','1','1','1','2','2','2','2','2','3','3','3'],
                 'type':['a','b','c','a','a','b','c','c','a','b','c','a']})

I use crosstab to produce a new dataframe and plot results:

pd.crosstab(test.cluster,test.type,normalize='index',margins=True).plot(kind='bar')

I would like to plot the row All as dotted horizontal benchmark lines of the same colour corresponding to each type to improve interpretation of the plot. Will appreciate help of this community!

标签： python-3.xpandasmatplotlibcrosstab

Answer 1

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
test = pd.DataFrame(
    {'cluster': ['1', '1', '1', '1', '2', '2', '2', '2', '2', '3', '3', '3'],
     'type': ['a', 'b', 'c', 'a', 'a', 'b', 'c', 'c', 'a', 'b', 'c', 'a']})
tab = pd.crosstab(test.cluster, test.type, normalize='index', margins=True)

fig, ax = plt.subplots()

# find the default colors
prop_cycle = plt.rcParams['axes.prop_cycle']
colors = prop_cycle.by_key()['color']

# make a bar plot using all rows but the last
tab.iloc[:-1].plot(ax=ax, kind='bar', color=colors)

# draw the horizontal dotted lines
for y, c in zip(tab.loc['All'], colors):
    ax.axhline(y=y, color=c, linestyle=':', alpha=0.5)
plt.show()