How to make multiple graphs with fig, ax

Question

I'm wondering how to make multiple bar graphs.

I made a dataframe with groupby like below

What I want is to show the min, max price of actual price and discount price each on separated two bar graph.

At first, I just tried to make graph using this code

df.plot(kind='bar', y='actual_price', stacked=False)

But this one show only one graph about the min and max of actual price.

I tried to use fig, ax = plt.subplot(1,2, figsize(10,10) as well. But it always gives the error

ax[0].bar(df.actual_price, stack=False)

How to solve this problem? How to know what the height is? Do I have to change the dataframe then?

Please share your opinion.

Thanks

Answer 1

Assuming a MultiIndex you might want to stack your dataset and use seaborn.catplot:

import seaborn as sns

sns.catplot(df.stack([0, 1])
              .rename_axis(['category', 'type', 'indicator'])
              .reset_index(name='value'),
            x='category', y='value', col='type', hue='indicator',
            kind='bar')

Output:

Or, using a loop:

import matplotlib.pyplot as plt

types = df.columns.get_level_values(0).unique()
# Index(['actual_price', 'discount_price'], dtype='object')

f, axes = plt.subplots(ncols=len(types))

for t, ax in zip(types, axes):
    ax.set_title(t)
    df[t].plot.bar(ax=ax)

Output:

reproducible input:

df = pd.DataFrame.from_dict({'index': ['Car', 'House', 'IT'],
                             'columns': [('actual_price', 'min'),
                              ('actual_price', 'max'),
                              ('discount_price', 'min'),
                              ('discount_price', 'max')],
                             'data': [[10, 100, 50, 50], [5, 8, 3, 4], [1, 7, 1, 5]],
                             'index_names': [None],
                             'column_names': [None, None],
                             }, orient='tight')