Plotly: How to hide axis titles in a plotly express figure with facets?

Question

Is there a simple way to hide the repeated axis titles in a faceted chart using plotly express? I tried setting

visible=True

In the code below, but that also hid the y axis tick labels (the values). Ideally, I would like to set hiding the repeated axes titles a default for faceted plots in general (or even better, just defaulting to showing a single x and y axis title for the entire faceted figure.

Here is the test code:

import pandas as pd
import numpy as np
import plotly.express as px
import string

# create a dataframe
cols = list(string.ascii_letters)
n = 50

df = pd.DataFrame({'Date': pd.date_range('2021-01-01', periods=n)})

# create data with vastly different ranges
for col in cols:
    start = np.random.choice([1, 10, 100, 1000, 100000])
    s = np.random.normal(loc=0, scale=0.01*start, size=n)
    df[col] = start + s.cumsum()

# melt data columns from wide to long
dfm = df.melt("Date")

fig = px.line(
    data_frame=dfm,
    x = 'Date',
    y = 'value',
    facet_col = 'variable',
    facet_col_wrap=6,
    facet_col_spacing=0.05,
    facet_row_spacing=0.035,
    height = 1000,
    width = 1000,
    title = 'Value vs. Date'
)

fig.update_yaxes(matches=None, showticklabels=True, visible=True)
fig.update_annotations(font=dict(size=16))
fig.for_each_annotation(lambda a: a.update(text=a.text.split("=")[-1]))

Final Code (accepted answer). Note plotly >= 4.9

import pandas as pd
import numpy as np
import plotly.express as px
import string
import plotly.graph_objects as go

# create a dataframe
cols = list(string.ascii_letters)
n = 50

df = pd.DataFrame({'Date': pd.date_range('2021-01-01', periods=n)})

# create data with vastly different ranges
for col in cols:
    start = np.random.choice([1, 10, 100, 1000, 100000])
    s = np.random.normal(loc=0, scale=0.01*start, size=n)
    df[col] = start + s.cumsum()

# melt data columns from wide to long
dfm = df.melt("Date")

fig = px.line(
    data_frame=dfm,
    x = 'Date',
    y = 'value',
    facet_col = 'variable',
    facet_col_wrap=6,
    facet_col_spacing=0.05,
    facet_row_spacing=0.035,
    height = 1000,
    width = 1000,
    title = 'Value vs. Date'
)

fig.update_yaxes(matches=None, showticklabels=True, visible=True)
fig.update_annotations(font=dict(size=16))
fig.for_each_annotation(lambda a: a.update(text=a.text.split("=")[-1]))

# hide subplot y-axis titles and x-axis titles
for axis in fig.layout:
    if type(fig.layout[axis]) == go.layout.YAxis:
        fig.layout[axis].title.text = ''
    if type(fig.layout[axis]) == go.layout.XAxis:
        fig.layout[axis].title.text = ''
        
# keep all other annotations and add single y-axis and x-axis title:
fig.update_layout(
    # keep the original annotations and add a list of new annotations:
    annotations = list(fig.layout.annotations) + 
    [go.layout.Annotation(
            x=-0.07,
            y=0.5,
            font=dict(
                size=16, color = 'blue'
            ),
            showarrow=False,
            text="single y-axis title",
            textangle=-90,
            xref="paper",
            yref="paper"
        )
    ] +
    [go.layout.Annotation(
            x=0.5,
            y=-0.08,
            font=dict(
                size=16, color = 'blue'
            ),
            showarrow=False,
            text="Dates",
            textangle=-0,
            xref="paper",
            yref="paper"
        )
    ]
)

fig.show()

Answer 1

This answer has five parts:

1. Hide subplot titles (not 100% sure you wanted to do that though...)
2. Hide y-axis tick values using fig.layout[axis].tickfont = dict(color = 'rgba(0,0,0,0)')
3. Set single axis labels using go.layout.Annotation(xref="paper", yref="paper")
4. the plotly figure
5. Complete code snippet at the end

One very important take-away here is that you can edit any element produced with a px function using plotly.graph_object references, like go.layout.XAxis.

---

1. Hide subplot titles

If you're otherwise happy with the way you've set up your fig, you can just include

for anno in fig['layout']['annotations']:
    anno['text']=''
    
fig.show()

2. Hide yaxis text

You can set the yaxis tickfont to transparent using the following in a loop

fig.layout[axis].tickfont = dict(color = 'rgba(0,0,0,0)')

That exact line is included in the snippet below that also removes y-axis title for every subplot.

3. Single axis labels

The removal of axis labels and inclusion of a single label requires a bit more work, but here's a very flexible setup that does exactly what you need and more if you'd like to edit your new labels in any way:

# hide subplot y-axis titles and x-axis titles
for axis in fig.layout:
    if type(fig.layout[axis]) == go.layout.YAxis:
        fig.layout[axis].title.text = ''
    if type(fig.layout[axis]) == go.layout.XAxis:
        fig.layout[axis].title.text = ''
        
# keep all other annotations and add single y-axis and x-axis title:
fig.update_layout(
    # keep the original annotations and add a list of new annotations:
    annotations = list(fig.layout.annotations) + 
    [go.layout.Annotation(
            x=-0.07,
            y=0.5,
            font=dict(
                size=16, color = 'blue'
            ),
            showarrow=False,
            text="single y-axis title",
            textangle=-90,
            xref="paper",
            yref="paper"
        )
    ] +
    [go.layout.Annotation(
            x=0.5,
            y=-0.08,
            font=dict(
                size=16, color = 'blue'
            ),
            showarrow=False,
            text="Dates",
            textangle=-0,
            xref="paper",
            yref="paper"
        )
    ]
)

fig.show()

4. Plot

5. Complete code:

import pandas as pd
import numpy as np
import plotly.express as px
import string
import plotly.graph_objects as go

# create a dataframe
cols = list(string.ascii_letters)
cols[0]='zzz'
n = 50

df = pd.DataFrame({'Date': pd.date_range('2021-01-01', periods=n)})

# create data with vastly different ranges
for col in cols:
    start = np.random.choice([1, 10, 100, 1000, 100000])
    s = np.random.normal(loc=0, scale=0.01*start, size=n)
    df[col] = start + s.cumsum()

# melt data columns from wide to long
dfm = df.melt("Date")

fig = px.line(
    data_frame=dfm,
    x = 'Date',
    y = 'value',
    facet_col = 'variable',
    facet_col_wrap=6,
    #facet_col_spacing=0.05,
    #facet_row_spacing=0.035,
    height = 1000,
    width = 1000,
    title = 'Value vs. Date'
)

fig.update_yaxes(matches=None, showticklabels=True, visible=True)
fig.update_annotations(font=dict(size=16))
fig.for_each_annotation(lambda a: a.update(text=a.text.split("=")[-1]))

# subplot titles
for anno in fig['layout']['annotations']:
    anno['text']=''

# hide subplot y-axis titles and x-axis titles
for axis in fig.layout:
    if type(fig.layout[axis]) == go.layout.YAxis:
        fig.layout[axis].title.text = ''
    if type(fig.layout[axis]) == go.layout.XAxis:
        fig.layout[axis].title.text = ''
        
# keep all other annotations and add single y-axis and x-axis title:
fig.update_layout(
    # keep the original annotations and add a list of new annotations:
    annotations = list(fig.layout.annotations) + 
    [go.layout.Annotation(
            x=-0.07,
            y=0.5,
            font=dict(
                size=16, color = 'blue'
            ),
            showarrow=False,
            text="single y-axis title",
            textangle=-90,
            xref="paper",
            yref="paper"
        )
    ] +
    [go.layout.Annotation(
            x=0.5,
            y=-0.08,
            font=dict(
                size=16, color = 'blue'
            ),
            showarrow=False,
            text="Dates",
            textangle=-0,
            xref="paper",
            yref="paper"
        )
    ]
)


fig.show()

Answer 2

As a side note on this, I found an even more direct way to eliminate the axis labels from within the plotly express call using the labels argument and providing it with a dict of labels with values of '' for the ones I wanted to eliminate.

This does not result in a single label at the overall figure level though, but if the figure title is descriptive enough of "Y vs. X", then maybe the lack of axis labels can be "excused"? ( or added in as @vestland demonstrated )

Note, you can "almost" eliminate the annoying repeated facet titles that have "=value" in every single sublot. i.e. if you add one more entry to the labels dict:

'variable': ''

then instead of getting "variable=variable level", you just get the the facet variable level, preceded by the "=" as in the chart below.

Full Code

import pandas as pd
import numpy as np
import plotly.express as px
import string

# create a dataframe
cols = list(string.ascii_letters)
n = 50

df = pd.DataFrame({'Date': pd.date_range('2021-01-01', periods=n)})

# create data with vastly different ranges
for col in cols:
    start = np.random.choice([1, 10, 100, 1000, 100000])
    s = np.random.normal(loc=0, scale=0.01*start, size=n)
    df[col] = start + s.cumsum()

# melt data columns from wide to long
dfm = df.melt("Date")

# make the plot
fig = px.line(
    data_frame=dfm,
    x = 'Date',
    y = 'value',
    facet_col = 'variable',
    facet_col_wrap=6,
    facet_col_spacing=0.05,
    facet_row_spacing=0.035,
    height = 1000,
    width = 1000,
    title = 'Value vs. Date',
    labels = {
        'Date': '',
        'value': '',
        'variable': ''
    }
)

# ensure that each chart has its own y rage and tick labels
fig.update_yaxes(matches=None, showticklabels=True, visible=True)

fig.show()