Plotly：通过子图连接图

Question

我有一个数据框，我可以绘制三种不同的图（条形图、堆积条形图和线图），但我只需要使用子图绘制一个图形。可以吗？

我的代码：

import pandas as pd
import plotly.express as px

data = pd.DataFrame([['A',0.25,0.25,0.25,0.25,125,0.032],
                     ['B',0.20,0.40,0.15,0.25,175,0.054],
                     ['C',0.25,0.25,0.30,0.20,192,0.065]])
data.columns = ['GROUP','Q1','Q2','Q3','Q4','SALES','IND']
data = data.set_index('GROUP')
data

fig1 = px.bar(data['SALES'], text_auto=True)
fig1.show()

fig2 = px.bar(data[['Q1','Q2','Q3','Q4']])
fig2.show()

fig3 = px.line(data['IND'])
fig3.show()type here

我只能添加两个图（条形图和线条图），但无法添加堆叠条形图

我试过这个：

from plotly.subplots import make_subplots
import plotly.graph_objects as go

fig = make_subplots(rows=2, cols=1)

df1 = pd.DataFrame(data['SALES'])
df1.reset_index(inplace=True)
fig.add_trace(go.Bar(x=df1['GROUP'],y=df1['SALES']),1,1)

df2 = pd.DataFrame(data['IND'])
df2.reset_index(inplace=True)
fig.add_trace(go.Scatter(x=df2['GROUP'],y=df2['IND']),2,1)

fig.show()

Answer 1

由于 x 轴是分类变量，因此条形图和堆叠条形图重叠。避免这种情况的第三种方法是数据框并将堆叠图的 x 轴更改为

[0,1,2]

和

[0.5,1.5,2.5]

以避免重叠。同时，第二个 y 轴用于读取值。最后，将第一个图的 x 轴刻度更改为两个图之间中点处的值和变量名称。可能有比我的方法更简单的方法。

from plotly.subplots import make_subplots
import plotly.graph_objects as go

fig = make_subplots(rows=2, cols=1, specs=[[{"secondary_y": True}],[{"secondary_y": False}]])

df1 = pd.DataFrame(data['SALES'])
df1.reset_index(inplace=True)
fig.add_trace(go.Bar(x=[0,1,2], y=df1['SALES'], width=0.48, name='SALES'),1,1)#df1['GROUP']

df3 = pd.DataFrame(data[['Q1','Q2','Q3','Q4']])
df3.reset_index(inplace=True)
fig.add_trace(go.Bar(x=[0.5,1.5,2.5], y=df3['Q1'], width=0.480, name='Q1'),1,1,secondary_y=True)#df3['GROUP']
fig.add_trace(go.Bar(x=[0.5,1.5,2.5], y=df3['Q2'], width=0.480, name='Q2'),1,1,secondary_y=True)
fig.add_trace(go.Bar(x=[0.5,1.5,2.5], y=df3['Q3'], width=0.480, name='Q3'),1,1,secondary_y=True)
fig.add_trace(go.Bar(x=[0.5,1.5,2.5], y=df3['Q4'], width=0.480, name='Q4'),1,1,secondary_y=True)
fig.update_layout(barmode='stack')
fig.update_xaxes(tickvals=[0.25,1.25,2.25], ticktext=['A','B','C'], row=1, col=1)

df2 = pd.DataFrame(data['IND'])
df2.reset_index(inplace=True)
fig.add_trace(go.Scatter(x=df2['GROUP'],y=df2['IND'], name='IND'),2,1)

fig.show()