hvPlot Bolded Grey Multi-index

[adam-morse]

Software Version Info

hvplot: 0.11.3

Description of expected behavior and the observed behavior

This is an issue with hvplot's bar rendering that I mentioned to the core developers in the 12/17/25 Holoviz Office Hours meeting.

The issue was the multi-index, primary index showing up in the dashboard with a bolded grey font. It was showing up fuzzy in our app and I was asked to fix it. But I couldn't figure it out. When I moved to Bokeh it was easily remedied.

My feelings aren't hurt if folks don't think its worth the time to fix or improve. Maybe moving to Bokeh to customize styling is the right path. In hindsight removing the label altogether through display(chart.opts(multi_level=False)) would have served my specific needs, but I didn't know it at the time.

Complete, minimal, self-contained example code that reproduces the issue

import pandas as pd
import numpy as np
import holoviews as hv
import hvplot.pandas 


# DataFrame: Scenario, Asset, Value, Value_formatted
data = [
    {"Scenario": "Default - High PV", "Asset": "PV", "Value": 120.0},
    {"Scenario": "Default - High PV", "Asset": "Battery", "Value": 300.0},
    {"Scenario": "Default - High PV", "Asset": "Gen-set", "Value": 50.0},
    {"Scenario": "Alt - No PV", "Asset": "PV", "Value": 0.0},
    {"Scenario": "Alt - No PV", "Asset": "Battery", "Value": 200.0},
    {"Scenario": "Alt - No PV", "Asset": "Gen-set", "Value": 120.0},
    {"Scenario": "Battery-Heavy", "Asset": "PV", "Value": 80.0},
    {"Scenario": "Battery-Heavy", "Asset": "Battery", "Value": 600.0},
    {"Scenario": "Battery-Heavy", "Asset": "Gen-set", "Value": 20.0},
]
df = pd.DataFrame(data)
unit_map = {"PV": "kWdc", "Battery": "kWh", "Gen-set": "kW"}
df['Value_formatted'] = df.apply(lambda r: f"{r['Value']:.1f} {unit_map[r['Asset']]}", axis=1)
# Preserve scenario display order
df['Scenario'] = pd.Categorical(df['Scenario'], categories=list(df['Scenario'].unique()), ordered=True)

chart = df.hvplot.bar(
    x="Scenario", y="Value", by="Asset",
    rot=45, hover=True, hover_cols=["Value_formatted"],
    title="Grouped bar: Scenario x Asset",
    responsive=True,
    height=300,
    xlabel="",
)

display(chart)

Stack traceback and/or browser JavaScript console output

Screenshots or screencasts of the bug in action

• I may be interested in making a pull request to address this

[maximlt]

Thanks for opening this issue @adam-morse!

To illustrate it:

The primary index appears in bold grey font and indeed looks somewhat fuzzy (likely due to its styling?).

I didn't know about multi_level! We could expose it for bar and barh. Otherwise, the backend_opts option is the "modern" way to customize a plot beyond what hvPlot/HoloViews both offer, but directly from an hvPlot call.

In the docs I've tried to describe how to use you. I think it deserves its own How-To guide as that's quite a common thing to do and there are some tricks that can help. For example, when you display a Bokeh model in a notebook you can actually click on the ... to expand it and inspect its properties and their values. This is how I found out about group_text_font_style and the other properties:

After clicking on ...: