如何基于给定DataFrame绘制100%堆叠柱状图并显示百分比？

Got it, let's turn your regular stacked bar chart into a 100% stacked version that shows percentage values directly on the bars. Here's a step-by-step solution tailored to your data:

Step 1: Calculate Grouped Sums & Normalize to Percentages #

First, we'll compute the total counts per name like you did initially, then normalize each value to its percentage of the group's total. This normalization is what creates the 100% stacked effect.

import pandas as pd
import matplotlib.pyplot as plt

# Your original DataFrame
df = pd.DataFrame({
    "name": ["Foo", "Foo", "Baar", "Foo", "Baar", "Foo", "Baar", "Baar"],
    "count_1": [5,10,12,15,20,25,30,35],
    "count_2": [100,150,100,25,250,300,400,500]
})

# 1. Group by name and sum the count columns
sum_df = df.groupby(['name'])[['count_1', 'count_2']].sum()

# 2. Convert sums to percentages (divide each row by its total, multiply by 100)
pct_df = sum_df.div(sum_df.sum(axis=1), axis=0) * 100

Step 2: Plot the 100% Stacked Bar Chart #

Use the percentage-based DataFrame to generate the stacked chart, and add basic styling for readability:

# Create the 100% stacked bar chart
ax = pct_df.plot(kind='bar', stacked=True, figsize=(8, 6), colormap='coolwarm')

# Add axis labels and title
ax.set_xlabel('Name')
ax.set_ylabel('Percentage (%)')
ax.set_title('100% Stacked Bar Chart: Count 1 vs Count 2 by Name')
ax.legend(title='Count Category')

Step 3: Add Percentage Labels to Each Bar Segment #

Loop through every segment of the stacked bars to place the percentage value dead-center on each section:

# Iterate over each bar segment to add labels
for p in ax.patches:
    # Grab the dimensions and position of the segment
    width, height = p.get_width(), p.get_height()
    x_pos, y_pos = p.get_xy()
    
    # Add formatted percentage text
    ax.text(
        x_pos + width/2,  # Center horizontally
        y_pos + height/2, # Center vertically
        f'{height:.1f}%',  # Show 1 decimal place for clarity
        ha='center',
        va='center'
    )

# Adjust layout to prevent label cutoff
plt.tight_layout()
plt.show()

Quick Explanation #

• Normalization: sum_df.sum(axis=1) calculates the total count per name, and div(..., axis=0) ensures each value is scaled relative to its group's total.
• Label Placement: Using ax.patches lets us target every individual segment of the stacked bars, so each percentage is placed exactly where it belongs.

内容的提问来源于stack exchange，提问作者Sean