Plotly：如何结合 make_subplots() 和 ff.create_distplot()？

Question

使用 plotly 创建多个子图既简单又优雅。 考虑以下示例，该示例并排绘制来自 dataframe 的两个系列：

Plot：

代码：

# imports
from plotly.subplots import make_subplots
import plotly.figure_factory as ff
import plotly.graph_objs as go
import pandas as pd
import numpy as np

# data
np.random.seed(123)
frame_rows = 40
n_plots = 6
#frame_columns = ['V_'+str(e) for e in list(range(1,n_plots+1))]
frame_columns = ['V_1', 'V_2']
df = pd.DataFrame(np.random.uniform(-10,10,size=(frame_rows, len(frame_columns))),
                  index=pd.date_range('1/1/2020', periods=frame_rows),
                    columns=frame_columns)
df=df.cumsum()+100
df.iloc[0]=100

# plotly setup
plot_rows=1
plot_cols=2
fig = make_subplots(rows=plot_rows, cols=plot_cols)

# plotly traces
fig.add_trace(go.Scatter(x=df.index, y=df['V_1']), row=1, col=1)
fig.add_trace(go.Scatter(x=df.index, y=df['V_2']), row=1, col=2)


fig.show()

用类似的对象替换go.Scatter() object 很容易：

Plot：

但我似乎找不到将此设置与ff.create_distplot()结合起来的方法：

分布图：

带有 distplot 的代码：

# imports
from plotly.subplots import make_subplots
import plotly.figure_factory as ff
import plotly.graph_objs as go
import pandas as pd
import numpy as np

# data
np.random.seed(123)
frame_rows = 40
n_plots = 6
#frame_columns = ['V_'+str(e) for e in list(range(1,n_plots+1))]
frame_columns = ['V_1', 'V_2']
df = pd.DataFrame(np.random.uniform(-10,10,size=(frame_rows, len(frame_columns))),
                  index=pd.date_range('1/1/2020', periods=frame_rows),
                    columns=frame_columns)
df=df.cumsum()+100
df.iloc[0]=100

# plotly setup
plot_rows=1
plot_cols=2
fig = make_subplots(rows=plot_rows, cols=plot_cols)

# plotly traces
fig.add_trace(go.Scatter(x=df.index, y=df['V_1']), row=1, col=1)
#fig.add_trace(go.Scatter(x=df.index, y=df['V_2']), row=1, col=2)

# distplot
hist_data = [df['V_1'].values, df['V_2'].values]
group_labels = ['Group 1', 'Group 2']
#fig2 = ff.create_distplot(hist_data, group_labels)

# combine make_subplots, go.Scatter and ff.create_distplot(
fig.add_trace(ff.create_distplot(hist_data, group_labels), row=1, col=2)

fig.show()

这会引发一个相当大的 ValueError。

原因似乎是go.Scatter()和ff.create_distplot()返回两种不同的数据类型； plotly.graph_objs.Scatter和plotly.graph_objs._figure.Figure ，分别。 并且它肯定似乎make_subplots不适用于后者。 或者有人知道解决这个问题的方法吗？

感谢您的任何建议！

Answer 1

事实证明，您不能直接执行此操作，因为make_subplots()不会直接接受plotly.graph_objs._figure.Figure object 作为add_trace()的参数。 但是您可以构建一个ff.create_distplot '并从该图中“窃取”数据并将它们应用到go.Histogram和go.Scatter()中接受的对象的组合中make_subplots() 。 你甚至可以用地毯/边距 plot 做同样的事情。

Plot：

代码：

# imports
from plotly.subplots import make_subplots
import plotly.figure_factory as ff
import plotly.graph_objs as go
import pandas as pd
import numpy as np

# data
np.random.seed(123)
frame_rows = 40
n_plots = 6
#frame_columns = ['V_'+str(e) for e in list(range(1,n_plots+1))]
frame_columns = ['V_1', 'V_2']
df = pd.DataFrame(np.random.uniform(-10,10,size=(frame_rows, len(frame_columns))),
                  index=pd.date_range('1/1/2020', periods=frame_rows),
                    columns=frame_columns)
df=df.cumsum()+100
df.iloc[0]=100

# plotly setup
plot_rows=2
plot_cols=2
fig = make_subplots(rows=plot_rows, cols=plot_cols)

# plotly traces
fig.add_trace(go.Scatter(x=df.index, y=df['V_1']), row=1, col=1)
fig.add_trace(go.Scatter(x=df.index, y=df['V_2']), row=2, col=1)

# distplot
hist_data = [df['V_1'].values, df['V_2'].values]
group_labels = ['Group 1', 'Group 2']
fig2 = ff.create_distplot(hist_data, group_labels)

fig.add_trace(go.Histogram(fig2['data'][0],
                           marker_color='blue'
                          ), row=1, col=2)

fig.add_trace(go.Histogram(fig2['data'][1],
                           marker_color='red'
                          ), row=1, col=2)

fig.add_trace(go.Scatter(fig2['data'][2],
                         line=dict(color='blue', width=0.5)
                        ), row=1, col=2)

fig.add_trace(go.Scatter(fig2['data'][3],
                         line=dict(color='red', width=0.5)
                        ), row=1, col=2)

# rug / margin plot to immitate ff.create_distplot
df['rug 1'] = 1.1
df['rug 2'] = 1
fig.add_trace(go.Scatter(x=df['V_1'], y = df['rug 1'],
                       mode = 'markers',
                       marker=dict(color = 'blue', symbol='line-ns-open')
                        ), row=2, col=2)

fig.add_trace(go.Scatter(x=df['V_2'], y = df['rug 2'],
                       mode = 'markers',
                       marker=dict(color = 'red', symbol='line-ns-open')
                        ), row=2, col=2)

# some manual adjustments on the rugplot
fig.update_yaxes(range=[0.95,1.15], tickfont=dict(color='rgba(0,0,0,0)', size=14), row=2, col=2)
fig.update_layout(showlegend=False)

fig.show()

Answer 2

对@vestland 解决方案的进一步建议是使用select_traces并迭代地组合您的新图形。 因此，您可以使用以下方法重新分配它们，而不是使用图中的数据重新绘制轨迹：

dist_fig = ff.create_distplot(hist_data=hist_data, group_labels=group_labels)

for trace in dist_fig.select_traces():
    fig.add_trace(trace, row=1, col=2)

这样您就不需要检查数据字段的确切索引值。 所以将新方法翻译成你的代码，它会是这样的：

from plotly.subplots import make_subplots
import plotly.figure_factory as ff
import plotly.graph_objs as go
import pandas as pd
import numpy as np

# data
np.random.seed(123)
frame_rows = 40
n_plots = 6
frame_columns = ['V_1', 'V_2']
df = pd.DataFrame(np.random.uniform(-10, 10, size=(frame_rows, len(frame_columns))),
                  index=pd.date_range('1/1/2020', periods=frame_rows),
                  columns=frame_columns)
df = df.cumsum() + 100
df.iloc[0] = 100

# plotly setup
plot_rows = 2
plot_cols = 2
fig = make_subplots(rows=plot_rows, cols=plot_cols)

# plotly traces
fig.add_trace(go.Scatter(x=df.index, y=df['V_1']), row=1, col=1)
fig.add_trace(go.Scatter(x=df.index, y=df['V_2']), row=2, col=1)

# distplot
hist_data = [df[col_name].to_list() for col_name in frame_columns]
dist_fig = ff.create_distplot(hist_data=hist_data, group_labels=frame_columns)

for trace in dist_fig.select_traces():
    fig.add_trace(trace, row=1, col=2)

df['rug 1'] = 1.1
df['rug 2'] = 1

fig.add_trace(go.Scatter(x=df['V_1'], y = df['rug 1'],
                       mode = 'markers',
                       marker=dict(color = 'blue', symbol='line-ns-open')
                        ), row=2, col=2)

fig.add_trace(go.Scatter(x=df['V_2'], y = df['rug 2'],
                       mode = 'markers',
                       marker=dict(color = 'orange', symbol='line-ns-open')
                        ), row=2, col=2)
fig.show()