Plotly龙卷风图X轴自定义：双镜像轴起始点统一设置

我用Plotly的make_subplot函数制作龙卷风图，采用Tableau的Sample Superstore数据集，当前图表中间存在空白区域，请问如何为两个镜像轴设置统一的起始点？

我的代码：

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

# 假设df是已加载的Sample Superstore数据集
df_oct = df.loc[(df['order_date'] > '2017-09-30') & (df['order_date'] < '2017-11-01') & (df['region'] == 'Central')]
df_sep = df.loc[(df['order_date'] > '2017-08-31') & (df['order_date'] < '2017-10-01') & (df['region'] == 'Central')]

states = pd.DataFrame(
            df.loc[(df['order_date'] > '2017-08-31') & (df['order_date'] < '2017-11-01') & (df['region'] == 'Central'), 'state'].unique()
            ,columns=['state'])

df_oct = states.merge(df_oct.groupby(by='state', as_index=False)['sales'].sum(), how='left', on='state')
df_sep = states.merge(df_sep.groupby(by='state', as_index=False)['sales'].sum(), how='left', on='state')

df_oct['month'] = 'October'
df_sep['month'] = 'September'
tornado_df = pd.concat([df_oct, df_sep])
tornado_df.fillna(0, inplace=True)

fig = make_subplots(
        rows=1
        ,cols=2
        ,vertical_spacing=0
)

fig_add = fig.add_trace(
            go.Histogram(
                x=tornado_df.loc[tornado_df['month'] == 'September', 'sales']
                ,y=tornado_df.loc[tornado_df['month'] == 'September', 'state']
                ,histfunc='sum'
                ,orientation='h'
                ,marker_color='#D95002'
                ,name='September'
                ,opacity=0.6)
            ,row=1
            ,col=1)

fig_add = fig.add_trace(
            go.Histogram(
                x=tornado_df.loc[tornado_df['month'] == 'October', 'sales']
                ,y=tornado_df.loc[tornado_df['month'] == 'October', 'state']
                ,histfunc='sum'
                ,orientation='h'
                ,marker_color='#523E89'
                ,name='October'
                ,opacity=0.6)
            ,row=1
            ,col=2)    

fig_add = fig.update_xaxes(
            autorange="reversed"
            ,row=1
            ,col=1)

fig_add = fig.update_yaxes(
            visible=False
            ,row=1
            ,col=2)

fig_add.show()

我的图表：

解决方法 #

要让两个镜像轴对齐零点、消除中间空白，需要做以下调整：

1. 计算统一的轴范围
   先找出两个月份销售额的最大值，以此作为左右轴的对称范围：

   max_sales = max(tornado_df['sales'].max(), abs(tornado_df['sales'].min()))
   

2. 设置固定轴范围并对齐零点
   对左右两个x轴分别指定范围，左轴范围设为[-max_sales, 0]（自带反转效果），右轴范围设为[0, max_sales]，同时关闭自动缩放：

   fig.update_xaxes(range=[-max_sales, 0], autorange=False, row=1, col=1)
   fig.update_xaxes(range=[0, max_sales], autorange=False, row=1, col=2)
   

3. 消除子图间距
   在make_subplots中添加horizontal_spacing=0，让两个子图紧贴在一起：

   fig = make_subplots(
       rows=1, cols=2,
       vertical_spacing=0,
       horizontal_spacing=0
   )
   

修改后的完整代码 #

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

# 假设df是已加载的Sample Superstore数据集
df_oct = df.loc[(df['order_date'] > '2017-09-30') & (df['order_date'] < '2017-11-01') & (df['region'] == 'Central')]
df_sep = df.loc[(df['order_date'] > '2017-08-31') & (df['order_date'] < '2017-10-01') & (df['region'] == 'Central')]

states = pd.DataFrame(
            df.loc[(df['order_date'] > '2017-08-31') & (df['order_date'] < '2017-11-01') & (df['region'] == 'Central'), 'state'].unique()
            ,columns=['state'])

df_oct = states.merge(df_oct.groupby(by='state', as_index=False)['sales'].sum(), how='left', on='state')
df_sep = states.merge(df_sep.groupby(by='state', as_index=False)['sales'].sum(), how='left', on='state')

df_oct['month'] = 'October'
df_sep['month'] = 'September'
tornado_df = pd.concat([df_oct, df_sep])
tornado_df.fillna(0, inplace=True)

# 计算最大销售额，用于对称轴范围
max_sales = max(tornado_df['sales'].max(), abs(tornado_df['sales'].min()))

fig = make_subplots(
        rows=1, cols=2,
        vertical_spacing=0,
        horizontal_spacing=0  # 消除子图间距
)

fig.add_trace(
            go.Histogram(
                x=tornado_df.loc[tornado_df['month'] == 'September', 'sales']
                ,y=tornado_df.loc[tornado_df['month'] == 'September', 'state']
                ,histfunc='sum'
                ,orientation='h'
                ,marker_color='#D95002'
                ,name='September'
                ,opacity=0.6)
            ,row=1
            ,col=1)

fig.add_trace(
            go.Histogram(
                x=tornado_df.loc[tornado_df['month'] == 'October', 'sales']
                ,y=tornado_df.loc[tornado_df['month'] == 'October', 'state']
                ,histfunc='sum'
                ,orientation='h'
                ,marker_color='#523E89'
                ,name='October'
                ,opacity=0.6)
            ,row=1
            ,col=2)    

# 设置统一的轴范围，对齐零点
fig.update_xaxes(range=[-max_sales, 0], autorange=False, row=1, col=1)
fig.update_xaxes(range=[0, max_sales], autorange=False, row=1, col=2)

fig.update_yaxes(visible=False, row=1, col=2)

fig.show()

这样修改后，左右轴的零点会完全对齐，中间空白区域消失，形成标准的龙卷风图样式。

内容的提问来源于stack exchange，提问作者Ars