[FIXED] sort values by sum after grouping pandas

Issue

hello i have new dataset by grouped. this is the result;

job            y
admin.         0    5227
               1    1045
blue-collar    0    5208
               1     517
entrepreneur   0     755
               1      96
housemaid      0     586
               1      82
management     0    1507
               1     255
retired        0     761
               1     331
self-employed  0     759
               1     111
services       0    2165
               1     260
student        0     364
               1     216
technician     0    3434
               1     589
unemployed     0     479
               1     109
unknown        0     166
               1      26

at this case, i want to plot into bar plot sort by sum of each job to get information top the most job, here the code i used for but it has a mistakes

import matplotlib.pyplot as plt
plt.figure(figsize=(6,6))
pekerjaan = df_new.groupby(['job','y'])['y'].size().unstack()
pekerjaan.sort_values(by='y',ascending=True).plot(kind='barh',stacked=True)
plt.title('Job')
plt.ylabel('Kind of job')
plt.xlabel('Total')
plt.show()

thank you in advance

Solution

Sample Data and Imports:

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

np.random.seed(25)

n = 100
df_new = pd.DataFrame({
    'job': np.random.choice(['admin', 'blue-collar', 'entrepreneur'],
                            p=[.4, .4, .2], size=n),
    'y': np.random.choice([0, 1], size=n)
})

Then sum across each row to get the row total, then sort by the row totals:

plt.figure(figsize=(6, 6))
plot_df = df_new.groupby(['job', 'y'])['y'].size().unstack()
plot_df['All'] = plot_df.sum(axis=1)
plot_df = plot_df.sort_values('All')

ax = plot_df.plot(kind='barh', y=[0, 1], stacked=True,
                  title='Job', xlabel='Kind of Job',
                  rot=0)
plt.tight_layout()
plt.show()

Summary Counts:

plot_df = df_new.groupby(['job', 'y'])['y'].size().unstack()

y              0   1
job                 
admin         19  17
blue-collar   24  25
entrepreneur  10   5

plot_df with the All column:

plot_df['All'] = plot_df.sum(axis=1)

y              0   1  All
job                      
admin         19  17   36
blue-collar   24  25   49
entrepreneur  10   5   15

After sort_values:

plot_df = plot_df.sort_values('All')

y              0   1  All
job                      
entrepreneur  10   5   15
admin         19  17   36
blue-collar   24  25   49

---

An alternative with crosstab + margins:

plt.figure(figsize=(6, 6))
plot_df = (
    pd.crosstab(df_new['job'], df_new['y'], margins=True)
        .iloc[:-1]
        .sort_values('All')
)

ax = plot_df.plot(kind='barh', y=[0, 1], stacked=True,
                  title='Job', xlabel='Kind of Job',
                  rot=0)
plt.tight_layout()
plt.show()

Both Produce:

Answered By - Henry Ecker