Python中比例堆叠条形图实现问题：现有代码如何优化堆叠效果？

你当前的问题是多次调用sns.barplot时，每个条形都从y=0开始绘制，导致重叠。要实现堆叠效果，只需要给后续的条形指定bottom参数，让它基于前面类别的总和来绘制。以下是修改后的完整代码：

# 导入库
import pandas as pd
import seaborn as sns
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches

# 读取数据
df = pd.read_csv('marsh fig 7.csv')
plt.rcParams["axes.labelsize"] = 12
my_dpi = 96  # 补充缺失的my_dpi定义
plt.figure(figsize=(600/my_dpi, 300/my_dpi), dpi=my_dpi)

# 计算每个类别的月度均值（sns.barplot默认绘制均值）
monthly_means = df.groupby('Month').mean().reset_index()
# 计算累积比例，作为堆叠的底部
monthly_means['C1_C2'] = monthly_means['Percent_C1'] + monthly_means['Percent_C2']
monthly_means['C1_C2_C3'] = monthly_means['C1_C2'] + monthly_means['Percent_C3']
monthly_means['C1_C2_C3_C4'] = monthly_means['C1_C2_C3'] + monthly_means['Percent_C4']
monthly_means['C1_C2_C3_C4_C5'] = monthly_means['C1_C2_C3_C4'] + monthly_means['Percent_C5']

# 绘制堆叠条形图
bar1 = sns.barplot(x="Month", y="Percent_C1", data=monthly_means, color='darkgreen')
bar2 = sns.barplot(x="Month", y="Percent_C2", data=monthly_means, color='mediumseagreen', bottom=monthly_means['Percent_C1'])
bar3 = sns.barplot(x="Month", y="Percent_C3", data=monthly_means, color='goldenrod', bottom=monthly_means['C1_C2'])
bar4 = sns.barplot(x="Month", y="Percent_C4", data=monthly_means, color='sienna', bottom=monthly_means['C1_C2_C3'])
bar5 = sns.barplot(x="Month", y="Percent_C5", data=monthly_means, color='steelblue', bottom=monthly_means['C1_C2_C3_C4'])
# 注意：原代码中C6的y参数名称错误，这里修正为Percent_C6（请根据实际列名调整）
bar6 = sns.barplot(x="Month", y="Percent_C6", data=monthly_means, color='darkblue', bottom=monthly_means['C1_C2_C3_C4_C5'])

# 添加图例
bar1_patch = mpatches.Patch(color='darkgreen', label='% C1')
bar2_patch = mpatches.Patch(color='mediumseagreen', label='% C2')
bar3_patch = mpatches.Patch(color='goldenrod', label='% C3')
bar4_patch = mpatches.Patch(color='sienna', label='% C4')
bar5_patch = mpatches.Patch(color='steelblue', label='% C5')
bar6_patch = mpatches.Patch(color='darkblue', label='% C6')

# 调整y轴范围为0-100（因为是比例总和100%）
plt.ylim(0, 100)
plt.text(0, 97, "Marsh", ha='left', va='top', fontsize=12, fontweight='normal', color="black")

# 双轴绘制水位线
ax2 = plt.twinx()
df2 = pd.read_csv('SRS1 Water Level.csv')
sns.pointplot(x="Month", y="Water Depth (cm)", data=df2, color="black", linestyles=':', label=None, errwidth=1, capsize=0.2, ax=ax2)

plt.legend(handles=[bar1_patch, bar2_patch, bar3_patch, bar4_patch, bar5_patch, bar6_patch])
plt.show()

关键修改说明： #

1. 计算累积底部值：先按月份聚合计算各类别的均值，再依次计算前n个类别的总和，作为下一个条形的bottom参数，确保每个条形堆叠在前一个之上。
2. 修正C6的列名：原代码中C6的y参数写的是"Component Proportion"，推测是笔误，需改为对应C6的列名（如Percent_C6），请根据实际数据集调整。
3. 调整y轴范围：原代码设置ylim(0,40)不符合比例总和100%的需求，改为(0,100)。
4. 补充缺失的my_dpi定义：原代码中未定义my_dpi，这里补充了默认值96，可根据需求调整。

内容的提问来源于stack exchange，提问作者msmit275