如何用Matplotlib实现Excel式的100%堆叠条形图绘制？

Absolutely! You can totally replicate a 100% stacked bar chart (the horizontal style, just like your Excel example) using Matplotlib and Pandas. The key fix you need is normalizing your data so each row sums to 1 (100%)—your current code makes a stacked bar chart, but not the 100% scaled version you're aiming for.

Step-by-Step Breakdown #

1. Normalize the Dataset: Divide every value in a row by the total sum of that row. This ensures each row adds up to 1, which translates directly to 100% in the chart.
2. Generate the 100% Stacked Chart: Use the normalized dataframe to create the horizontal stacked bar plot.
3. Polish the Visual: Tweak labels, legends, and add optional percentage annotations to make the chart easy to read.

Modified Working Code #

import pandas as pd
import matplotlib.pyplot as plt

# Your original dataset
data = [
    [0.4, 0.3, 0.2, 0.1],
    [0.5, 0.3, 0.6, 0.1],
    [0.1, 0.4, 0.2, 0.8],
]

columns = ["A", "B", "C", "D"]
df = pd.DataFrame(data=data, columns=columns, index=["Empty", "Wrong", "Correct"])

# Normalize data so each row sums to 1 (100%)
df_normalized = df.div(df.sum(axis=1), axis=0)

# Create the 100% stacked horizontal bar chart
ax = df_normalized.plot(kind="barh", stacked=True, figsize=(10, 6))

# Add percentage labels to each bar segment
for p in ax.patches:
    width = p.get_width()
    if width > 0:  # Skip segments with no value
        ax.text(
            p.get_x() + width / 2,
            p.get_y() + p.get_height() / 2,
            f"{width:.1%}",
            ha="center",
            va="center",
            fontsize=10
        )

# Customize chart elements
plt.ylabel("Category")
plt.xlabel("Percentage")
plt.title("100% Stacked Horizontal Bar Chart")
plt.legend(title="Groups", bbox_to_anchor=(1.05, 1), loc="upper left")

plt.tight_layout()
plt.show()

# Print normalized data for reference
print("Normalized Data (each row sums to 1):")
print(df_normalized)

Quick Explanations #

• The line df.div(df.sum(axis=1), axis=0) handles normalization: axis=1 calculates the sum of each row, and axis=0 ensures we divide each value by its row's total.
• Adding percentage annotations makes it instantly clear what proportion each segment takes up.
• Moving the legend to the side prevents overlap with the chart, which boosts readability.

内容的提问来源于stack exchange，提问作者Salvo Dragotta