使用绘图小面网格绘制多条垂直线

Question

我想绘制 2 条垂直线来显示两种不同的平均值。使用图面网格再次将其分为不同的子图。

在下面，我在单独的子图中显示

cut

。在每个子图中，我有两种不同的颜色表示

val

。对于每个子图，我想显示 val 和 cut 的平均值。但我目前只显示剪切的平均值。

import seaborn as sns
import plotly.express as px

diamonds = sns.load_dataset('diamonds')

diamonds['val'] = np.random.randint(1, 3, diamonds.shape[0])

grpval = diamonds.groupby(['cut','val'])['price'].mean()
print(grpval)


fig = px.histogram(data_frame=diamonds, 
                   x='price', 
                   facet_col='cut',
                   color = "val", 
                   facet_col_wrap = 2,
                   )

for c,idx in zip(diamonds['cut'].unique(),[(1,1),(1,2),(2,1),(2,2),(3,1)]):
    df = diamonds[diamonds['cut'] == c]
    fig.add_vline(x=df['price'].tail(1).values[0], line_width=1, line_dash='solid', line_color='red', row=idx[0], col=idx[1])

fig.show()

我尝试绘制两种方法，但只得到一种。

Answer 1

以下是向绘图添加两条垂直平均线的方法：

import numpy as np
import seaborn as sns
import plotly.express as px

diamonds = sns.load_dataset("diamonds")

diamonds["val"] = np.random.randint(1, 3, diamonds.shape[0])

grpval = diamonds.groupby(["cut", "val"])["price"].mean()
# reshape the mean values into cut and val dimensions
grpval_per_cut = np.reshape(grpval.values, (diamonds["cut"].unique().size, -1))

# add color sequence to have same colors on histogram and lines
color_discrete_sequence = px.colors.qualitative.G10

fig = px.histogram(
    data_frame=diamonds,
    x="price",
    facet_col="cut",
    color="val",
    facet_col_wrap=2,
    color_discrete_sequence=color_discrete_sequence,
)


# iterate over mean values and subplots (I could not figure out a direct way
# to associate the subplot with the correct mean value, so I arranged it "by hand"
for mean_per_val, (row, col) in zip(
    grpval_per_cut, [(3, 1), (3, 2), (2, 2), (2, 1), (1, 1)]
):
    for val, line_color in zip(mean_per_val, color_discrete_sequence):
        fig.add_vline(
            x=val,
            line_width=1,
            line_dash="solid",
            line_color=line_color,
            row=row,
            col=col,
        )

fig.show()