Pandas boxplot: set color and properties for box, median, mean

Question

I have a DataFrame with a MultiIndex:

# -*- coding: utf-8 -*-
import numpy as np
import pandas as pd

# dataframe with dates
dates = pd.DataFrame()
dates[\'2         


        

Answer 1

Try seaborn

# Box Plot
import seaborn as sns
%matplotlib inline
sns.boxplot(data=data['fixed acidity'])
plt.show()

Answer 2

Before your bp.set_xlabel("") statement, try this instead:

p = plt.gca()
p.set_xlabel("")
p.set_title("Some plot", fontsize=60)
p.tick_params(axis='y', labelsize=60)
p.tick_params(axis='x', labelsize=60)

Answer 3

Screenpavers answer worked well.

Here's a complete example:

# -*- coding: utf-8 -*-
import numpy as np
import pandas as pd

# dataframe with dates
dates = pd.DataFrame()
dates['2016'] = pd.date_range(start='2016', periods=4, freq='60Min')
dates['2017'] = pd.date_range(start='2017', periods=4, freq='60Min')
dates['2018'] = pd.date_range(start='2018', periods=4, freq='60Min')
dates.reset_index()
dates = dates.unstack()

# multi-indexed dataframe
df = pd.DataFrame(np.random.randn(36, 3))
df['concept'] = np.repeat(np.repeat(['A', 'B', 'C'], 3), 4)
df['datetime'] = pd.concat([dates, dates, dates], ignore_index=True)
df.set_index(['concept', 'datetime'], inplace=True)
df.sort_index(inplace=True)
df.columns = ['V1', 'V2', 'V3']
df.info()


# demonstrate how to customize the display different elements:
boxprops = dict(linestyle='-', linewidth=4, color='k')
medianprops = dict(linestyle='-', linewidth=4, color='k')

bp = df.boxplot(column=['V1'],
                by=df.index.get_level_values('datetime').year,
                showfliers=False, showmeans=True,
                boxprops=boxprops, medianprops=medianprops,
                return_type='dict')

# boxplot style adjustments
[[item.set_linewidth(4) for item in bp[key]['boxes']] for key in bp.keys()]
[[item.set_linewidth(4) for item in bp[key]['fliers']] for key in bp.keys()]
[[item.set_linewidth(4) for item in bp[key]['medians']] for key in bp.keys()]
[[item.set_linewidth(4) for item in bp[key]['means']] for key in bp.keys()]
[[item.set_linewidth(4) for item in bp[key]['whiskers']] for key in bp.keys()]
[[item.set_linewidth(4) for item in bp[key]['caps']] for key in bp.keys()]

[[item.set_color('g') for item in bp[key]['boxes']] for key in bp.keys()]
# seems to have no effect
[[item.set_color('b') for item in bp[key]['fliers']] for key in bp.keys()]
[[item.set_color('m') for item in bp[key]['medians']] for key in bp.keys()]
[[item.set_markerfacecolor('k') for item in bp[key]['means']] for key in bp.keys()]
[[item.set_color('c') for item in bp[key]['whiskers']] for key in bp.keys()]
[[item.set_color('y') for item in bp[key]['caps']] for key in bp.keys()]

# get rid of "boxplot grouped by" title
plt.suptitle("")

# label adjustment
p = plt.gca()
p.set_xlabel("")
p.set_title("Some plot", fontsize=30)
p.tick_params(axis='y', labelsize=30)
p.tick_params(axis='x', labelsize=30)

returns:

Answer 4

I just found another solution to plot with much less code directly from pandas (without having to manipulate the matplotlib-object afterwards):

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt


df = pd.DataFrame(np.random.rand(10, 5), columns=['A', 'B', 'C', 'D', 'E'])
ax = df.plot(kind='box',
             color=dict(boxes='r', whiskers='r', medians='r', caps='r'),
             boxprops=dict(linestyle='-', linewidth=1.5),
             flierprops=dict(linestyle='-', linewidth=1.5),
             medianprops=dict(linestyle='-', linewidth=1.5),
             whiskerprops=dict(linestyle='-', linewidth=1.5),
             capprops=dict(linestyle='-', linewidth=1.5),
             showfliers=False, grid=True, rot=0)
ax.set_xlabel('Foo')
ax.set_ylabel('Bar in X')
plt.show()

yields:

The only thing I haven't figured out is how to adjust the color of the means when showmeans=True. But in most cases this should be fine..

Hope it helps!