在同一图上绘制多个变量，并按站 ID 绘制面板图

Question

我有 2 个时间序列数据帧，它们来自两个二维数组。这些数据帧的结构是：

生成示例数据帧

import pandas as pd
import numpy as np
date_range = pd.period_range('1981-01-01','1981-01-04',freq='D')
x = np.arange(8).reshape((4,2))
y = np.arange(8).reshape((4,2))
x = pd.DataFrame(x, index = date_range, columns = ['station1','station2'])
y = pd.DataFrame(y, index = date_range, columns = ['station1','station2'])
print(x)
          station1 station2
1981-01-01  0        1
1981-01-02  2        3
1981-01-03  4        5
1981-01-04  6        7

目标

我想生成一个多图，其中 'x' 和 'y' 的值在同一个图上绘制为线，x 和 y 按颜色分割，但每个站有多个图的“行” .使用上面的示例代码，每个单独的图表将绘制不同的站列。

我尝试过的

我尝试了 seaborn 路线：首先将两个数据帧连接在一起 - 每个 df 代表一个变量，因此我将它们添加为键以在连接后命名这些变量。然后我使用了 melt 来对它们进行多重绘图：

df = pd.concat([x , y], keys = ['Var1', 'Var2'])
meltdf = df.melt(var_name = 'Station', value_name = 'Value', ignore_index = False)
print(meltdf)
                  Station  Value
Var1 1981-01-01  station1      0
     1981-01-02  station1      2
     1981-01-03  station1      4
     1981-01-04  station1      6
Var2 1981-01-01  station1      0
     1981-01-02  station1      2
     1981-01-03  station1      4
     1981-01-04  station1      6
Var1 1981-01-01  station2      1
     1981-01-02  station2      3
     1981-01-03  station2      5
     1981-01-04  station2      7
Var2 1981-01-01  station2      1
     1981-01-02  station2      3
     1981-01-03  station2      5
     1981-01-04  station2      7

我想将 Var1 和 Var2 的值绘制为 station1 的同一图表上的线，station2 的相同，依此类推。我想保留日期作为索引，因为这些应该是时间序列图，“日期”沿 x 轴。我试过这个 non-working 代码（例如）：

import seaborn as sns
sns.relplot(data=df, x = 'Var1', y = 'Var2', kind = 'line', hue = 'keys', row = 'Station')

我应该“双重融化” dfs 以将变量类型作为自己的 col 吗？ concat + keys 步骤似乎不正确。

Answer 1

pd.concat 和 pd.melt，你在正确的轨道上，其次是 seaborn relplot。我会这样处理它：

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

#data generation
import numpy as np
np.random.seed(123)
date_range = pd.period_range('1981-01-01','1981-01-04',freq='D')
x = np.random.randint(1, 10, (4,2))
y = np.random.randint(1, 10, (4,2))
x = pd.DataFrame(x, index = date_range, columns = ['station1','station2'])
y = pd.DataFrame(y, index = date_range + pd.to_timedelta(1, unit="D"), columns = ['station1','station2'])

#keep information where each data point comes from
x["key"], y["key"] = "x", "y"

#combining dataframes and reshaping 
df = pd.concat([x, y]).melt(["key"], var_name="station", value_name="station_value",  ignore_index = False)

#plotting - the datetime conversion might not be necessary 
#depending on the datetime format of your original dataframes
#best approach is conversion to datetime index when creating the dataframes
fg = sns.relplot(data=df, x = pd.to_datetime(df.index.to_timestamp()), y = "station_value", kind = "line", hue = "key", row = "station")

#shouldn't be necessary but this example had too many ticks for the interval
from matplotlib.dates import DateFormatter, DayLocator
fg.axes[0,0].xaxis.set_major_locator(DayLocator(interval=1))
fg.axes[0,0].xaxis.set_major_formatter(DateFormatter("%y-%m-%d"))

plt.show()

示例输出：

如果pandas版本不能处理重复的索引条目，我们可以重写为：

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

#data generation
import numpy as np
np.random.seed(123)
date_range = pd.period_range('1981-01-01','1981-01-04',freq='D')
x = np.random.randint(1, 10, (4,2))
y = np.random.randint(1, 10, (4,2))
x = pd.DataFrame(x, index = date_range, columns = ['station1','station2'])
y = pd.DataFrame(y, index = date_range + pd.to_timedelta(1, unit="D"), columns = ['station1','station2'])

#keep information where each data point comes from
x["key"], y["key"] = "x", "y"
#moving index into a column 
x = x.reset_index()
y = y.reset_index()
#and changing it to datetime values that seaborn can understand
#only necessary because your example contains pd.Period data
x["index"] = pd.to_datetime(x["index"].astype(str))
y["index"] = pd.to_datetime(y["index"].astype(str))

#combining dataframes and reshaping 
df = pd.concat([x, y]).melt(["index", "key"], var_name="station", value_name="station_value")

#plotting
fg = sns.relplot(data=df, x = "index", y = "station_value", kind = "line", hue = "key", row = "station")

#shouldn't be necessary but this example had too many ticks for the interval
from matplotlib.dates import DateFormatter, DayLocator
fg.axes[0,0].xaxis.set_major_locator(DayLocator(interval=1))
fg.axes[0,0].xaxis.set_major_formatter(DateFormatter("%y-%m-%d"))

plt.show()