Adding another widget to a heatmap plot when using groupby


I’ve been trying to add a second widget to a heatmap plot - a discrete slider that will allow to set the maximum value of the clim where I’m using the convenience of the groupby parameter to also create a widget to select the subset of data.

For example below gives me the base plot I’d like:

import pandas as pd
import hvplot.pandas
import numpy as np
import panel as pn
from bokeh.sampledata import sea_surface_temperature as sst

df = sst.sea_surface_temperature

df['sea'] = 'sea1'

df2 = df.copy()
df2['sea'] = 'sea2'
df2_temps = df2['temperature'].to_numpy()
rng = np.random.default_rng()
df2['temperature'] = df2_temps

df_plot = pd.concat([df, df2])

plot = df_plot.hvplot.heatmap(x='time.month', y='', C='temperature', 
                       height=500, width=500, groupby=['sea'])

I would then like to add a DiscreteSlider such as:

max_clim = pn.widgets.DiscreteSlider(name='clim', value=20, options=list(range(1, 21)))

in order to set the color scale max as in:

df_plot.hvplot.heatmap(x='time.month', y='', C='temperature', 
                       height=500, width=500, groupby=['sea'], clim=(0, max_clim))

I’ve tried various forms of the information here: Widgets — hvPlot 0.8.2 documentation to make the connection including a callback function but I just cannot get it to work.

I can add the widget to the widget box and things like that but I have a feeling I’m misunderstanding something in how the groupby widget and connection is made behind the scenes when using the groupby option. Is it possible to ‘add’ a widget in this situation?

You may be able to use .interactive for this.

At this section in the docs for .interactive, a widget is used to set a plotting parameter.

1 Like

Thanks very much for your reply @droumis.

I was able to achieve what I wanted using the interactive framework with the code I’ve pasted below. I hadn’t used the interactive bit before so that was interesting to discover!

If I don’t get the group like I do below and instead use groupby in the heatmap plot then it errors with Nesting a DynamicMap inside a DynamicMap is not supported.

As I’m essentially writing the call back anyway to get the grouping I think I will construct it all myself using either this method or the usual @pn.depends instead of the groupby.

grouped = df_plot.groupby('sea')

def get_group(select):
    return grouped.get_group(select)

select = pn.widgets.Select(name='group_select', options=df_plot['sea'].unique().tolist(), value='sea1')

interactive_plot = hvplot.bind(get_group, select).interactive()

clim_test = pn.widgets.DiscreteSlider(name='clim', value=20, options=list(range(1, 31)))

interactive_plot.hvplot.heatmap(x='time.month', y='', C='temperature', 
                       height=500, width=500, clim=(0, clim_test))