Agfdhyk

Can someone help me that why plt.tight_layout failed to tight the layout of the last nest_pie chart?

plt.tight_layout() has applied to every figure except last one. So strange it seems to me that plt.show() can show every figure, but .tight_layout() can not tight everyone.

code is here:

def all_pie_nested():

for i in a:

    fig, ax = plt.subplots()

    data0 = df.groupby(i)['income'].sum()

    data0.plot.pie(autopct='%.1f%%')

    ax.set(aspect=1)

    for i1 in a:

        if i1 != i:

            size = 0.4



            fig, ax = plt.subplots()



            data1 = df.groupby([i, i1])['income'].sum()

            data0.plot.pie(ax=ax, radius=1 - size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

            data1.plot.pie(ax=ax, radius=1, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

            ax.set(aspect=1)



            for i2 in a:

                if i2 != i1 and i2 != i:

                    fig, ax = plt.subplots()



                    data2 = df.groupby([i, i1, i2])['income'].sum()



                    data0.plot.pie(ax=ax, radius=1-size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                    data1.plot.pie(ax=ax, radius=1, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                    data2.plot.pie(ax=ax, radius=1+size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                    ax.set(aspect=1)

    plt.tight_layout()

    plt.show()

How many figures do you want ? One or multiple ? If one, why do you call subplots multiple times ? If multiple, you may rather want to call tight_layout() specifically for each figure inside the loops:

fig.tight_layout()

Just like the picture below: Figure_3 & Figure_5 come from the same code , however, they are different:

for i2 in a:

    if i2 != i1 and i2 != i:

    fig, ax = plt.subplots()



    data2 = df.groupby([i, i1, i2])['income'].sum()



    data0.plot.pie(ax=ax, radius=1-size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

    data1.plot.pie(ax=ax, radius=1, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

    data2.plot.pie(ax=ax, radius=1+size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

    ax.set(aspect=1)

I've tried to call plt.tight_layout() inside each loop, and set the figure.autolayout rcParam to True under for i in a:, neither of them can make Figure_5 with same features as
Figure_3.

.....................parting line........................

The only sulotion, for the moment, is setting size to 0.3 from 0.4, and plt.rcParams['figure.autolayout'] = True before `for i in a:

Hope someone can explain the machanism of it later.

def all_pie_nested():



    plt.rcParams['figure.autolayout'] = True # replace call plt.tight_layout()



    for i in a:

        fig, ax = plt.subplots()    



        data0 = df.groupby(i)['income'].sum()

        data0.plot.pie(autopct='%.1f%%')

        ax.set(aspect=1)

        for i1 in a:

            if i1 != i:

                size = 0.3           # change from size = 0.4



                fig, ax = plt.subplots()



                data1 = df.groupby([i, i1])['income'].sum()

                data0.plot.pie(ax=ax, radius=1 - size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                data1.plot.pie(ax=ax, radius=1, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                ax.set(aspect=1)



                for i2 in a:

                    if i2 != i1 and i2 != i:

                        fig, ax = plt.subplots()



                        data2 = df.groupby([i, i1, i2])['income'].sum()

                        data0.plot.pie(ax=ax, radius=1-size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                        data1.plot.pie(ax=ax, radius=1, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                        data2.plot.pie(ax=ax, radius=1+size, autopct='%.1f%%', wedgeprops=dict(width=size, edgecolor='w'))

                        ax.set(aspect=1)

        plt.show()