It sounds simple but I cant get a stacked bar chart. I got a series after grouping by a column in a dataframe.
s1 = df_sort.groupby('loan_status')['count_loans'].sum()
s1
loan_status
Fully paid 8045
Not fully paid 1533
Name: count_loans, dtype: int64
I have tried
s1.T.plot.bar(stacked=True)
s1.plot(kind="bar", stacked=True)
which produces two individual bars?
thanks.
CodePudding user response:
You should change the groupby line to:
df = df.groupby(by = 'loan_status').sum().T
Then you can plot with:
df.plot(kind = 'bar', stacked = True)
plt.show()
CodePudding user response:
You have to convert your Series
to a DataFrame
else you can't transpose:
s1.to_frame().T.plot(kind='bar', stacked=True, rot=0)
plt.show()