Colors in python bar plot when using df.barplot

Question

I am able to get very nice colorful bar plots when I use matplotlibs commands. This was very helpful which uses some coloring commands like

 for box in bp['boxes']:
 # change outline color
 box.set( color='#7570b3', linewidth=2)
 # change fill color
 box.set( facecolor = '#1b9e77' )

However, the problem in using these boxplot commands, data should be grouped already as you want on the x-axis. My last question on S.O. was on that and as suggested in the reply, I can just use

df.boxplot('data1',by='data2')

This works out great for me as I can use my data as it is. However, I cannot use some coloring options as above here. In fact, parameter color is also not accepted as an argument like as shown in several examples like this

df.plot(kind='box',color='red')

but I still prefer using the df.boxplot commands (may be I need to learn more on other options).

Clearly, I am new to python and even newer to plotting tools so I do not understand how each of these options work in depth but would like to get insights on what is the quickest way of making a bar plot and adding colors to it.

SO my question is how do I add colors to my box plot if I use df.boxplot

Edit: Adding working example to show what I am looking for

if I have a dataframe as this

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

df=pd.DataFrame(np.random.rand(10,2),columns=['data1','data2'])
df['group']=['A','B','C','A','B','C','A','B','C','A']

and I am using following command to draw a box plot where data is grouped by column group and I want to see the spread of column `data1' in that.

plt.figure(3)    
df.boxplot(column='data1',by='group')

How do I add colors to it?

I know how to add colors if I use this command

 plt.figure(1)
 data2plot=[df['data1'],df['data2']]
 plt.boxplot(data2plot)

This does not give the plot which I want and to get the plot which I want, I really have to use the following code:

 grA=[val for ind, val in zip(df['group'],df['data1']) if ind=='A']
 grB=[val for ind, val in zip(df['group'],df['data1']) if ind=='B']
 grC=[val for ind, val in zip(df['group'],df['data1']) if ind=='C']

 plt.figure(2)
 data2plot_new=[grA,grB,grC]
 plt.boxplot(data2plot_new)

I know how to add colors if I use the above method but it does not look efficient.

Another option I know where 'color' parameter can be used is this

 df.plot(data2plot,kind='box',color='red')

but again even here I need to do some data transformation to get the plot I want.

So again the question is how do I change plot properties like color, thickness etc. if i use the first option.

I hope this is clear.

PagMax · Accepted Answer

More complete answer from @Goyo should have been

box1=df.boxplot('data1',by='group',return_type='dict',patch_artist=False)
[x.set(color='g',linewidth=2) for x in box1['data1']['boxes']] 
[x.set(facecolor='r') for x in box1['data1']['boxes']]

(@Goyo, while your answer surely helped me in digging for more information, it was interesting that you were complaining about an incomplete question from me and then posted an answer which was incomplete! ).

I am posting this as an answer because it serves my current purpose.

Colors in python bar plot when using df.barplot

Answers (2)

Related Questions