如何使用seaborn和twinx在条形图上绘制直线图

Question

我正试图建立一个阴谋，在那里我可以根据空气中的化学物质检查鸟类的死亡率。我所面临的挑战是如何将情节相互叠加

下面是我写的代码。基本上给定了6种化学物质，我有六个独立的轴，在每个轴上，我以周数为x轴，绘制出所述化学物质与鸟类死亡率的关系

即使图形的来源来自一个数据帧（X轴没有区别），Seaborn也会分别绘制这两个图形

我尝试使用bird_chem_df.plot()和vanilla matplotlib来查看这是否仅限于Seaborn，但没有运气

很想知道到底是什么原因导致X轴不一致，尽管数据来自同一来源，并且它们共享相同的自变量

如果你想玩数据=https://drive.google.com/file/d/1KKkSoy3xQno_vE_-LSoviClmqC3UDcW6/view?usp=sharing

# Import the CSV in the link above
# bird_chem_df = pd.read_csv('...')

# Using `target_chemicals_chart` as our array, we will loop through each
# and plot it against bird mortality
target_chemicals_chart = ['Toluene', 'o-Xylene', 'm,p-Xylene','Ethylbenzene', 'Benzene', 'PM2.5']
target_chemicals_display = {
    'Toluene':[0,2.5],
    'o-Xylene':[0,2.5],
    'm,p-Xylene':[0,2.5],
    'Benzene':[0,2.5],
    'Ethylbenzene':[0,8], 
    'PM2.5':[0,25]
}
chemical_index = 0

# Setup the plot
fig, axs = plt.subplots(2,3,figsize=(12,7))
fig.subplots_adjust(left=7, bottom=7, right=9, top=9, wspace=0.2, hspace=0.5)

# Plot each axis
for row in axs:
    for col in row:
        target_chemical = target_chemicals_chart[chemical_index]
        col.set_title(f"{target_chemical} vs. Bird Mortality")
        sb.barplot(ax=col, x="week_number", y=target_chemical, data=bird_chem_df, ci=None, color='lightsteelblue')
        col.set_ylim(target_chemicals_display[target_chemical])
        bird = col.twinx()
        sb.lineplot(ax=bird, x="week_number", y="total_bird_deaths", data=bird_chem_df, color='red')
        col.set_xlabel("Weeks since Spill")
        col.set_ylabel("Average Result for Chemical - ug/m3")

        # Increase font size due to fig configurations
        for item in ([col.title, col.xaxis.label, col.yaxis.label] +
             col.get_xticklabels() + bird.get_yticklabels() + col.get_yticklabels()):
            item.set_fontsize(20)

        chemical_index += 1

Answer 1

• 条形图记号的位置为0索引，线形图记号的位置不为0索引
• 将多维数组axs转换为1-D，以便于访问和迭代
• 在这种情况下，条形图的x轴为'week_number'，与线形图相同，因此两个图之间的记号位置数量相同。因此，不是使用x='week_number'打印，而是通过指定x=ax.get_xticks()将线图打印到与条形图相同的勾号位置
• 另一个选项是对两个绘图使用x=bird_chem_df.index（因为索引是从0开始的RangeIndex），然后将xtick标签更改为bird_chem_df.week_number
• 在python 3.8.11、pandas 1.3.2、matplotlib 3.4.3、seaborn 0.11.2

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

data = {'week_number': [19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34], 'Benzene': [0.62, 0.459, 0.542, 0.631, 0.56, 0.61, 0.691, 1.109, 0.524, 0.87, 0.896, 0.665, 0.898, 0.862, 0.611, 0.37], 'Ethylbenzene': [0.364, 0.204, 0.372, 0.36, 0.225, 0.412, 0.332, 0.659, 0.241, 1.7, np.nan, 1.2, 7.35, 0.352, 0.267, 0.154], 'PM2.5': [14.621, 12.561, 11.174, 18.307, 11.285, 20.202, 16.111, 13.057, 11.301, 12.214, 18.173, 21.308, 14.009, 14.111, 9.805, 7.818], 'Toluene': [1.339, 0.999, 1.18, 2.019, 1.217, 1.797, 1.478, 1.896, 1.552, 2.201, 1.101, 1.416, 1.215, 1.598, 1.356, 0.877], 'm,p-Xylene': [1.059, 0.842, 1.186, 1.116, 0.693, 1.372, 0.972, 2.103, 0.568, 1.783, 1.736, 1.486, 1.983, 1.082, 0.909, 0.354], 'o-Xylene': [0.525, 0.328, 0.356, 0.409, 0.265, 0.396, 0.32, 0.634, 0.266, 1.543, 0.74, 2.0, 0.93, 0.374, 0.328, 0.15], 'total_bird_deaths': [13, 14, 224, 87, 120, 165, 224, 252, 303, 416, 580, 537, 592, 713, 442, 798]}
bird_chem_df = pd.DataFrame(data)

target_chemicals_chart = ['Toluene', 'o-Xylene', 'm,p-Xylene','Ethylbenzene', 'Benzene', 'PM2.5']
target_chemicals_display = {'Toluene': [0, 2.5], 'o-Xylene': [0, 2.5], 'm,p-Xylene': [0, 2.5], 'Benzene': [0, 2.5], 'Ethylbenzene': [0, 8], 'PM2.5': [0, 25]}
chemical_index = 0

# Setup the plot
fig, axs = plt.subplots(2, 3, figsize=(15, 7))

# convert axs to a 1-D array
axs = axs.ravel()

for ax in axs:
    target_chemical = target_chemicals_chart[chemical_index]
    ax.set_title(f"{target_chemical} vs. Bird Mortality")
    p1 = sns.barplot(ax=ax, x="week_number", y=target_chemical, data=bird_chem_df, ci=None, color='lightsteelblue')

    ax.set_ylim(target_chemicals_display[target_chemical])
    bird = ax.twinx()

    # plot against the same tick values as the bar plot, with x=ax.get_xticks()
    p2 = sns.lineplot(ax=bird, x=ax.get_xticks(), y="total_bird_deaths", data=bird_chem_df, color='red', marker='o')

    ax.set_xlabel("Weeks since Spill")
    ax.set_ylabel("Average Result for Chemical - ug/m3")
   
fig.tight_layout()

• 绘制所有化学物质的列

data = {'week_number': [19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34], 'Benzene': [0.62, 0.459, 0.542, 0.631, 0.56, 0.61, 0.691, 1.109, 0.524, 0.87, 0.896, 0.665, 0.898, 0.862, 0.611, 0.37], 'Ethylbenzene': [0.364, 0.204, 0.372, 0.36, 0.225, 0.412, 0.332, 0.659, 0.241, 1.7, np.nan, 1.2, 7.35, 0.352, 0.267, 0.154], 'PM2.5': [14.621, 12.561, 11.174, 18.307, 11.285, 20.202, 16.111, 13.057, 11.301, 12.214, 18.173, 21.308, 14.009, 14.111, 9.805, 7.818], 'Toluene': [1.339, 0.999, 1.18, 2.019, 1.217, 1.797, 1.478, 1.896, 1.552, 2.201, 1.101, 1.416, 1.215, 1.598, 1.356, 0.877], 'm,p-Xylene': [1.059, 0.842, 1.186, 1.116, 0.693, 1.372, 0.972, 2.103, 0.568, 1.783, 1.736, 1.486, 1.983, 1.082, 0.909, 0.354], 'o-Xylene': [0.525, 0.328, 0.356, 0.409, 0.265, 0.396, 0.32, 0.634, 0.266, 1.543, 0.74, 2.0, 0.93, 0.374, 0.328, 0.15], 'total_bird_deaths': [13, 14, 224, 87, 120, 165, 224, 252, 303, 416, 580, 537, 592, 713, 442, 798]}
bird_chem_df = pd.DataFrame(data)

target_chemicals_display = {'Toluene': [0, 2.5], 'o-Xylene': [0, 2.5], 'm,p-Xylene': [0, 2.5], 'Benzene': [0, 2.5], 'Ethylbenzene': [0, 8], 'PM2.5': [0, 25]}
target_chemicals_chart = ['Toluene', 'o-Xylene', 'm,p-Xylene','Ethylbenzene', 'Benzene', 'PM2.5']

for target_chemical in target_chemicals_display:
    # Setup the plot
    fig, axs = plt.subplots(2, 3, figsize=(15, 7))
    axs = axs.ravel()

    for ax in axs:
        ax.set_title(f"{target_chemical} vs. Bird Mortality")

        p1 = sns.barplot(ax=ax, x=bird_chem_df.index, y=target_chemical, data=bird_chem_df, ci=None, color='lightsteelblue')

        ax.set_ylim(target_chemicals_display[target_chemical])
        bird = ax.twinx()
        p2 = sns.lineplot(ax=bird, x=bird_chem_df.index, y="total_bird_deaths", data=bird_chem_df, color='red', marker='o')

        # set the x-axis tick label to be the week numbers
        ax.set_xticks(ax.get_xticks())
        ax.set_xticklabels(bird_chem_df.week_number)

        ax.set_xlabel("Weeks since Spill")
        ax.set_ylabel("Average Result for Chemical - ug/m3")


    fig.tight_layout()
    plt.show()