Changing the y label for histogram in plotly express does not work

[vikigenius]

Hi I am using plotly express to plot a histogram.

import plotly.express as px

px.histogram(
    cdf,
    x='diag',
    histnorm='probability',
    labels={'diag': 'cosine similarity'},
    title='Cosine Similarity of Diagonal Samples',
    color_discrete_sequence=['green'],
)

The y label is shown as count, I want the y label to be called probability.
I tried passing labels={'diag': 'cosine similarity', 'count': 'probability'} but that doesn't work? How do I change the y label ?

[set92]

Keeps happening in 4.14.3

I fixed it a little bit with yaxis_title but in the popup that appear when you put the mouse over the bars still shows "count".

fig = px.histogram(
    n_orders, labels={"value": "# Pedidos", "count": "# Clientes"}
).update_layout(
    title={"text": "Numero de pedidos por cliente", "x": 0.5}, yaxis_title="# Clientes"
)
fig.show()

[mmonto95]

Keeps happening

[zhenliu2012]

Still happening with 5.6.0

[kaoutaar]

you need to change it using
fig.update_layout(yaxis_title="probability")

[nicolaskruchten]

This works fine for me for both the y axis title and the hoverlabel... I'm not sure what the problem is.

[brylie]

@nicolaskruchten, I am experiencing this issue today.

In our project, all texts need to be localized, so we cannot rely on auto-generated labels such as "sum of" or "probability".

The following code produces the correct chart, but the aggregate category label is prefixed with "sum of" in both the axis label and hover text.

px.histogram(
    chart_data,
    x="role_name",
    y="total_minutes",
    color="work_type",
    labels={
        # notice here the use of _("") localization function
        "role_name": _("Caregiver role"),
        "total_minutes": _("total minutes"),
        "work_type": _("Type of work"),
    },
)

The desired output above would be that the y-axis label and hover text would just be "total minutes" rather than "sum of total minutes", the latter of which we cannot localize.

[brylie]

Given the ambiguity of the original request, I will gladly open a second issue related to changing the y-axis label and hover text, if needed.

[nicolaskruchten]

@brylie this is a bit of a different issue but I may as well respond here so folks scrolling through find it ;)

There's no set of arguments you can pass to PX to localize these "sum of" and "probability" strings unfortunately, so you'll have to mutate them explicitly with something like the code below. For completeness, you'll need to localize not only the field names as you have done above, but also the values within the colums ("Sat", "Sun" in my example, and "Nurse" in your example above), for which there is no Plotly functionality, and you'll need to do in the underlying data.

import plotly.express as px

fig = px.histogram(px.data.tips(), x="day", y="total_bill")
fig.for_each_trace(lambda t: t.update(hovertemplate=t.hovertemplate.replace("sum of", "somme de")))
fig.for_each_yaxis(lambda a: a.update(title_text=a.title.text.replace("sum of", "somme de")))

[brylie]

@nicolaskruchten, thank you for your quick response and examples. The example seems like a good temporary work-around while exploring the underlying issue of prepending English text to chart axis labels and hover texts.

In my case, I'd rather the text weren't prepended at all, since my axis label ("total minutes") is correct.