Mastering Plotly and Dash Data Visualization with Python/

...

/

Advanced Subplots

Advanced Subplots

Build on your knowledge of subplots to undertake advanced customization and formatting.

We'll cover the following...

Unevenly distributed plots

What if we want our plots to take up unequal space on our main figure?

Uneven column widths

It turns out that by passing a list to column_widths inside make_subplots, we can have our subplots take up unequal space. The values are normalized internally but be sure to have the column widths add to one for maximum code interpretability. Here we have set 65% (0.65) of the total width for the first subplot and 35% (0.35) for the second subplot. This is seen in the next code cell.

Uneven row widths

Instead of column_widths, we can use row_heights if we stack subplots vertically. We demonstrate this in the code cell after the next one.

Press + to interact
# Import libraries
import pandas as pd
import numpy as np
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
# Import dataset
activations = pd.read_csv('/usr/local/csvfiles/activations.csv')
# Make subplots
# Uneven columns width
fig = make_subplots(rows=1, 
                    cols=2, 
                    subplot_titles=['Mish', 'Softplus'],
                    shared_yaxes=True,
                    column_widths=[0.65, 0.35])
# Create traces and add to figure
trace1 = go.Scatter(x=activations['x'], 
                    y=activations['mish'], 
                    name='Mish',
                    mode='lines')
trace2 = go.Scatter(x=activations['x'], 
                    y=activations['softplus'], 
                    name='Softplus',
                    mode='lines')
fig.add_trace(trace1, row=1, col=1)
fig.add_trace(trace2, row=1, col=2)
# Show
fig.show()

We could also have uneven row heights. Let’s look at an example and how we deal with this:

Press + to interact
# Import libraries
import pandas as pd
import numpy as np
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
# Import dataset
activations = pd.read_csv('/usr/local/csvfiles/activations.csv')
# Make subplots
# Uneven row height
fig = make_subplots(rows=2, 
                    cols=1, 
                    subplot_titles=['Mish', 'Softplus'],
                    shared_xaxes=True,
                    row_heights=[0.65, 0.35])
# Create traces and add to figure
trace1 = go.Scatter(x=activations['x'], 
                    y=activations['mish'], 
                    name='Mish',
                    mode='lines')
trace2 = go.Scatter(x=activations['x'], 
                    y=activations['softplus'], 
                    name='Softplus',
                    mode='lines')
fig.add_trace(trace1, row=1, col=1)
fig.add_trace(trace2, row=2, col=1)
# Show
fig.show()

Advanced grid configurations

Suppose we want certain subplots to take up more than others within a large grid. Here we take advantage of the specs argument when creating Plotly subplots. The specs argument allows you to pass in a list of lists where each inner list represents a subplot grid row and the elements inside an inner list correspond to subplot grid columns. Note that inside these inner lists we can either pass in a dictionary or None. We now demonstrate these ideas ...