How do i get pie chart labels to line up with the correct values from dataframe?-CodePudding

My pie chart is coming out okay, the segments are the correct sizes, but the labels aren't on the correct places.

# declaring exploding pie
explode = [0, 0.1, 0, 0, 0,0]
# define Seaborn color palette to use
palette_color = sns.color_palette('pastel')
  
# plotting data on chart
fig=plt.pie(combined_df.groupby(['Continent'])['total_consumption'].sum(), colors=palette_color,labels=combined_df['Continent'].unique(),
        explode=explode, autopct='%.0f%%', labeldistance=0.9,)
plt.show

the values that should be plotted

Continent
Africa           227.0
Asia             128.7
Europe           431.9
North America    147.5
Oceania           42.5
South America     83.2

CodePudding user response：

My suspicion is that the order of the returned results of unique() and value_counts() is not consistent.

value_counts() returned is order by count, and unique() is order by appearance in the array.

You can try this:

value_cnts = combined_df['Continent'].value_counts()
fig = plt.pie(value_cnts, colors=palette_color, labels=value_cnts.index, 
    explode=explode, autopct='%.0f%%', labeldistance=0.9)

CodePudding user response：

Try to not sort values on value_counts:

fig=plt.pie(combined_df['Continent'].value_counts(sort=False),  # <- do not sort
            colors=palette_color,
            labels=combined_df['Continent'].unique(),  # because unique is not sorted
            explode=explode, autopct='%.0f%%', labeldistance=0.9,)

Demo:

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

palette_color = sns.color_palette('pastel')
explode = [0, 0.1, 0, 0, 0,0]

np.random.seed(2023)
data = data = np.random.choice(
    ['Africa', 'Asia', 'Europe', 'North America', 'Oceania', 'South America'],
    p=(0.213989, 0.121324, 0.407146, 0.139046, 0.040064, 0.078431),
    size=1000
)
combined_df = pd.DataFrame({'Continent': data})

fig = plt.pie(combined_df['Continent'].value_counts(sort=False), 
              autopct='%.0f%%', labeldistance=0.9, 
              labels=combined_df['Continent'].unique(),
              colors=palette_color, explode=explode)
plt.show()