How to rotate the axis labels in plotly scatterplotmatrix?

Question

With plotly express scattermatrix

import plotly.express as px
df = px.data.iris()
fig = px.scatter_matrix(df)
fig.show()

I would like to rotate the axis-labels (not the tick-labels) of the y-axis, which correspond to the columns in the dataframe, so that they are horizontal. Because, for longer names and more subplots the labels start overlapping pretty quickly. Is that currently doable?

Answer 1

• axis ticks can be rotated. I cannot find any way to rotate labels
• as a workaround, replace labels and add a key

import plotly.express as px

df = px.data.iris()
fig = px.scatter_matrix(df)

# rotate text in yaxes, if text
fig.update_layout(
    {
        f"yaxis{n+1 if n>0 else ''}": {"tickangle": 45}
        for n in range(len(fig.data[0].dimensions))
        if isinstance(fig.data[0].dimensions[n].values[0], str)
    }
)

# get labels of dimensions in splom
labels = [d.label for d in fig.data[0].dimensions]

# replace labels in splom with identifier
fig.update_traces(
    dimensions=[d.update(label=chr(ord("a")+n)) for n, d in enumerate(fig.data[0].dimensions)]
)

# add a key for labels
for n, l in enumerate(labels):
    fig.add_annotation(
        x=(n) / (len(labels)-1),
        y=1.2,
        xref="paper",
        yref="paper",
        text=f"{chr(ord('a')+n)}. {l}",
        align="center",
        showarrow=False,
    )

fig