如何向多个子图添加附加图

Posted 2023-03-12

【中文标题】如何向多个子图添加附加图

【英文标题】：How to add an additional plot to multiple subplots

【发布时间】：2021-12-28 19:06:36

【问题描述】：

我想生成一对线图，其中一个用作基准。

我可以用下面的代码生成这样的图。

但是，我希望我可以有 6 对地块，以阿尔汉格尔斯克作为每个地块的基准线。例如，其中一个将是这样的：

.

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

data = 'year': ['1998','1998','1998','1998','1998','1998','1998','1999','1999','1999','1999','1999','1999','1999'],'region':['Adygea','Altai Krai','Amur Oblast','Arkhangelsk','Astrakhan','Bashkortostan','Belgorod','Adygea','Altai Krai','Amur Oblast','Arkhangelsk','Astrakhan','Bashkortostan','Belgorod'],  'sales':[8.8, 19.2,21.2, 10.6,18,17.5,23, 10, 17.8, 20.5, 12.6, 19.9, 16, 21]

df1 = pd.DataFrame(data)

plt.figure(figsize=(12, 6))

palette1 = c:'#079b51' if c=='Astrakhan' else 'grey' for c in df1['region'].unique()

sns.lineplot(x= 'year', y='sales', data=df1,hue='region', palette=palette1) # or sns.relplot
plt.legend(bbox_to_anchor=(1.05, 1), loc='upper left')

我的代码是可重现的。

我尝试了以下方法，但显然不起作用。我不知道如何遍历我的 6 个区域中的每一个来比较 Astkrakhan 线图。它可能应该包含不等于/等于 'Astrakhan' 之类的条件？谢谢。

fig, axs = plt.subplots(2,3, figsize=(18,15))
for i in enumerate(df1.region.unique()):
    sns.lineplot(x= 'year', y='sales', data=df1,ax = axs[i])

这会引发 ValueError：无法解释参数 Adygea 的值 y。

df2 = df1.pivot(index='year', columns='region', values='sales') # converting the rows into columns
df_a = df2[['Arkhangelsk']]
df_r = df2.loc[:, ~df2.columns.isin(['Arkhangelsk'])] ## all other columns

fig, axes = plt.subplots(2, 3)
for col, ax in zip(df2.columns, axes.ravel()):
    sns.lineplot(x = "year", y = col, data = df_a, ax = ax, linestyle="--")
    sns.lineplot(x = "year", y = col, data = df_r, ax = ax)

【参考方案1】：

使用pandas.DataFrmame.plot，与seaborn一样，使用matplotlib

# convert the year column to an int
df.year = df.year.astype(int)

# pivot the data to a wide format
dfp = df.pivot(index='year', columns='region', values='sales')

# get the columns to plot and compare
compare = 'Arkhangelsk'
cols = dfp.columns.tolist()
cols.remove(compare)

# set color dict
color = c:'#079b51' if c=='Arkhangelsk' else 'grey' for c in df['region'].unique()

# plot the data with subplots
axes = dfp.plot(y=cols, subplots=True, layout=(2, 3), figsize=(16, 10), sharey=True, xticks=dfp.index, color=color)

# flatten the array
axes = axes.flat  # .ravel() and .flatten() also work

# extract the figure object
fig = axes[0].get_figure()

# iterate through each axes
for ax in axes:
    
    # plot the comparison column
    dfp.plot(y=compare, ax=ax, color=color)
    
    # adjust the legend if desired
    ax.legend(loc='upper center', bbox_to_anchor=(0.5, 1.10), frameon=False, ncol=2)
    
fig.suptitle('My Plots', fontsize=22, y=0.95)
plt.show()