df_plot = df.groupby(['class', 'survived']).size().reset_index().pivot(columns='class', index='survived', values=0)

class     First  Second  Third
survived                      
0            80      97    372
1           136      87    119