How to rearrange and set color in a nested pie plot

Question

I am trying to draw a graph like this:

The dataset is:

    Console Type    Company Release Units_sold
0   PlayStation 2   Home    Sony    2000    155000000
1   Nintendo DS Handheld    Nintendo    2004    154002000
2   Nintendo Switch Hybrid  Nintendo    2017    122550000
3   Game Boy    Handheld    Nintendo    1989    118690000
4   PlayStation 4   Home    Sony    2013    117200000
5   PlayStation Home    Sony    1994    102490000
6   Wii Home    Nintendo    2006    101630000
7   PlayStation 3   Home    Sony    2006    87400000
8   Xbox 360    Home    Microsoft   2005    84000000
9   PlayStation Portable    Handheld    Sony    2004    82000000
10  Game Boy Advance    Handheld    Nintendo    2001    81510000
11  Nintendo 3DS    Handheld    Nintendo    2011    75940000
12  NES Home    Nintendo    1983    61910000
13  Xbox One    Home    Microsoft   2013    58500000
14  SNES    Home    Nintendo    1990    49100000
15  Nintendo 64 Home    Nintendo    1996    32930000
16  PlayStation 5   Home    Sony    2020    32100000
17  Xbox    Home    Microsoft   2001    24000000
18  GameCube    Home    Nintendo    2001    21740000
19  Xbox Series X/S Home    Microsoft   2020    18500000
20  PlayStation Vita    Handheld    Sony    2011    15000000
21  Wii U   Home    Nintendo    2012    13560000
22  SNES Classic    Dedicated   Nintendo    2017    5280000
23  NES Classic Dedicated   Nintendo    2016    3560000
24  Color TV-Game   Dedicated   Nintendo    1977    3000000

And this is my code:

df_3_outer = df_3.sort_values('Company').reset_index()
# only keep the console and units_sold columns
df_3_outer = df_3_outer[['Console', 'Units_sold']]

df_3_inner = df_3.groupby(['Company'])['Units_sold'].sum().reset_index()
fig, ax = plt.subplots(figsize=(9,6))

inner_colors = ['#156EAF', '#DB2018', '#56B45B']
outer_colors =['#5599CC', '#EA6727', '#83C143']

outer_colors[labels.index('')]
size = 0.8
r = 2

ax.pie(df_3_outer['Units_sold'], labels=df_3_outer.Console, radius=r, colors=outer_colors, 

       wedgeprops=dict(width=size, edgecolor='w'))



ax.pie(df_3_inner.Units_sold, labels=df_3_inner.Company, radius=r-size, colors=inner_colors,
                     labeldistance=0.6,
                     textprops=dict(color="w", fontsize=15),
       )

However, it gives the plot like this:

As you can see, it is not complete.

How can I arrange the outer ring to match the inner ring, and how can I filter the color pattern like the top figure?

Answer 1

You need to repeat the colors for the outer ring by the number of elements in each. Pandas' groupby(...).count() can be used to count, and np.repeat to create an array with the repetitions.

If you sort both on company name and units sold, the outer ring will take that order into account.

As some consoles have a very small number of units sold, the names will overlap and clutter the plot. You may want to filter them away.

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

data = [['PlayStation 2', 'Home', 'Sony', 2000, 155000000], ['Nintendo DS', 'Handheld', 'Nintendo', 2004, 154002000], ['Nintendo Switch', 'Hybrid', 'Nintendo', 2017, 122550000], ['Game Boy', 'Handheld', 'Nintendo', 1989, 118690000], ['PlayStation 4', 'Home', 'Sony', 2013, 117200000], ['PlayStation', 'Home', 'Sony', 1994, 102490000], ['Wii', 'Home', 'Nintendo', 2006, 101630000], ['PlayStation 3', 'Home', 'Sony', 2006, 87400000], ['Xbox 360', 'Home', 'Microsoft', 2005, 84000000], ['PlayStation Portable', 'Handheld', 'Sony', 2004, 82000000], ['Game Boy Advance', 'Handheld', 'Nintendo', 2001, 81510000], ['Nintendo 3DS', 'Handheld', 'Nintendo', 2011, 75940000], ['NES', 'Home', 'Nintendo', 1983, 61910000], ['Xbox One', 'Home', 'Microsoft', 2013, 58500000], ['SNES', 'Home', 'Nintendo', 1990, 49100000], ['Nintendo 64', 'Home', 'Nintendo', 1996, 32930000], ['PlayStation 5', 'Home', 'Sony', 2020, 32100000], ['Xbox', 'Home', 'Microsoft', 2001, 24000000], ['GameCube', 'Home', 'Nintendo', 2001, 21740000], ['Xbox Series X/S', 'Home', 'Microsoft', 2020, 18500000], ['PlayStation Vita', 'Handheld', 'Sony', 2011, 15000000], ['Wii U', 'Home', 'Nintendo', 2012, 13560000], ['SNES Classic', 'Dedicated', 'Nintendo', 2017, 5280000], ['NES Classic', 'Dedicated', 'Nintendo', 2016, 3560000]]
df_3 = pd.DataFrame(data=data, columns=['Console', 'Type', 'Company', 'Release', 'Units_sold'])

df_3 = df_3[df_3['Units_sold'] > 10000000] # optionally filter away very small values
df_3_outer = df_3.sort_values(['Company', 'Units_sold'])
df_3_outer = df_3_outer[['Console', 'Units_sold']]
df_3_inner = df_3.groupby(['Company'])['Units_sold'].sum().reset_index()
df_3_counts = df_3.groupby(['Company'])['Units_sold'].count()

fig, ax = plt.subplots(figsize=(12, 9))

inner_colors = ['#156EAF', '#DB2018', '#56B45B']
outer_colors_single = ['#5599CC', '#EA6727', '#83C143']
outer_colors = np.repeat(outer_colors_single, df_3_counts)
size = 0.3
r = 1

ax.pie(df_3_outer['Units_sold'], labels=df_3_outer['Console'], radius=r, colors=outer_colors,
       wedgeprops=dict(width=size, edgecolor='w'))
ax.pie(df_3_inner['Units_sold'], labels=df_3_inner['Company'], radius=r - size, colors=inner_colors,
       labeldistance=0.6,
       textprops=dict(color='w', fontsize=15))
plt.tight_layout()
plt.show()