Pandas groupby two columns and plot

Question

I have a dataframe like this:

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline
df = pd.DataFrame({'category': list('XYZXY'), 'B': range(5,10),'sex': list('mfmff')})

I want to plot count of sex male or female based on category from column 'category'.

I tried:
df.groupby(['category','sex'])['category','sex'].count().plot.bar()

But this gives:

How do I get the count of sex per category?

Solution 1

Various Methods of Groupby Plots

Data

import numpy as np
import pandas as pd
df = pd.DataFrame({'category': list('XYZXY'),
                   'sex': list('mfmff'),
                   'ThisColumnIsNotUsed': range(5,10)})
df
category sex ThisColumnIsNotUsed
0   X   m   5
1   Y   f   6
2   Z   m   7
3   X   f   8
4   Y   f   9

Using crosstab

pd.crosstab(df['category'],df['sex']).plot.bar()

Using groupby+unstack:

(df.groupby(['sex','category'])
   .count().unstack('sex').plot.bar())

Using pivot_table:

pd.pivot_table(df,index = 'category',
               columns = 'sex',aggfunc ='count').plot.bar()

Using seaborn:

import seaborn as sns
sns.countplot(data=df,x='category',hue='sex')
or,
sns.catplot(data=df,kind='count',x='category',hue='sex')

output

Solution 2

IIUC,

df.groupby(['category','sex']).B.count().unstack().reset_index()\
.plot.bar(x = 'category', y = ['f', 'm'])

Edit: If you have multiple columns, you can use groupby, count and droplevel.

new_df = df.groupby(['category','sex']).count().unstack()
new_df.columns = new_df.columns.droplevel()
new_df.reset_index().plot.bar()

Solution 3

You can also use this

pd.pivot_table(df, values = 'B', index = 'category', columns = 'sex',
               aggfunc = lambda x: len(x)).plot.bar()

which results in exactly the same plot.

Comments

[Vaishali]

@ScottBoston, haha. Great minds?;) Wish you a very happy new year!

[Scott Boston]

@Vaishali Happy 2019!

[BhishanPoudel]

Happy new year to you all guys, also small followup, what to do if I have more than three columns, like B, C, D, etc. ? Can we choose any of the other column, or the chosen column must have unique values ?

[Vaishali]

@astro123, pl see the edit. And happy new year to you:)

[Jason Goal]

everything run smoothly but the figure does not show up, why?

[Vaishali]

@Jason Goal, You need plt.show() at the end of the code

[kirti purohit]

Hello , I have a similar problem it would be really helpful if you could please help me out with this -StackOverflow