How to create grouped barplot from dataframe grouped by two columns with intervals

Question

I grouped the data by two columns ( 'a' and 'b' ). And for each column, I cut it into bins. For each crossed bin, I would like to have the sum of 'c' . Then I want to create a grouped plot for the sum of 'c' . However, when I use .plot(kind='bar') after the grouped data, the bars were not grouped together.

grouped = df.groupby([pd.cut(df['a'], [0, 1, 2, 3]), pd.cut(df['b'], [0, 101, 300, 500])])
grouped = grouped.c.sum()
grouped.plot(kind='bar')

ungrouped barplot

I would like the graph to be grouped by 'a' with different color standing for 'b' . I have tried .unstack().plot(kind='bar') ; however, it returns an error:

cannot determine next label for type

Answer 1

Given the following example data

df = pd.DataFrame({'a': [8, 49, 81, 52, 22, 24, 32, 67, 24, 21, 78, 16, 86, 19, 4, 88, 4, 20, 20, 1],
                   'b': [2, 49, 49, 70, 31, 47, 98, 26, 55, 36, 17, 39, 56, 80, 52, 36, 42, 69, 73, 70],
                   'c': [22, 99, 31, 95, 16, 32, 81, 20, 58, 23, 53, 5, 0, 81, 8, 68, 16, 36, 35, 54]})

and your setup

grouped = df.groupby([pd.cut(df['a'], [16, 30, 67, 86]),
                      pd.cut(df['b'], [0, 26, 49, 100])])
grouped = grouped['c'].sum()

you have two options

•  (grouped.reset_index() .pivot('a', 'b') .plot(kind='bar')); 

  

•  grouped = grouped.unstack() grouped.columns = grouped.columns.categories grouped.plot(kind='bar');