How to control heatmap colorbar ticks?

hvPlot
1

I would like to have a heatmap which meets the following requirements:

  1. Specify limits and colors within the limits.
  2. Ticks on the colorbar should be the limits specified above and depend on the data. For example, even if the data lies in [5, 10], if I specify a limits as [0, 100], then the colorbar should extend from 0-100.

Expected behavior

image
Of course in the above image, the colors will be linearly scaled.

Observed behavior

The above screenshot, but instead of the red 200.

The code for the panel is:

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

hv.extension("bokeh")

liststr_col = [f"col_{ind}" for ind in range(10)]
liststr_row = [f"row_{ind}" for ind in range(10)]
listfloat_quantities = np.arange(1, 101, 1).astype(float).reshape(10, 10)
listfloat_quantities[3, 4] = np.nan
df = pd.DataFrame(listfloat_quantities
    , columns=liststr_col, index=liststr_row
)

listfloat_color_levels = [0, 11, 21, 51, 75, 200]
liststr_cmap = ["#79de79", "#a8e4ef", "#fcfc99", "#fb6962", "#ffffff"]

dict_heatmapopts = dict(color_levels=listfloat_color_levels, cmap=liststr_cmap)

heatmap = df.hvplot.heatmap(shared_axes=False).opts(
    title=str(listfloat_color_levels),
    **dict_heatmapopts,
)
hvdim_value = hv.Dimension("value", value_format=lambda x: "%.3f" % x)
label = hv.Labels(heatmap, vdims=hvdim_value).opts(
    padding=0, text_color="black", text_alpha=1.0, text_font_size="11pt"
)
listplots = []
listplots.append(heatmap * label)
# listplots.append(heatmap * label)
gc.collect()
hv.Layout(listplots).cols(2)
2

I think you are looking for clim in heatmap like:

...
heatmap = df.hvplot.heatmap(shared_axes=False).opts(
    title=str(listfloat_color_levels), clim=(0, 200),
    **dict_heatmapopts,
)
...

image

1 Like
3

Thank you very much @Hoxbro - that worked. I guess I left this part out of my question earlier: is it possible to control the xticks in the colorbar? I would like for the xticks to be at the levels I specify in listfloat_color_levels.

4

@Hoxbro Nevermind, I figured it out. I ended up following python - How do I manually set the tick locations of a colorbar for a Points plot in HoloViews? - Stack Overflow and doing the following

liststr_col = [f"col_{ind}" for ind in range(10)]
liststr_row = [f"row_{ind}" for ind in range(10)]
listfloat_quantities = np.arange(1, 101, 1).astype(float).reshape(10, 10)
listfloat_quantities[3, 4] = np.nan
listfloat_quantities[3, 5] = 150
df = pd.DataFrame(listfloat_quantities, columns=liststr_col, index=liststr_row)

listfloat_color_levels = [0, 11, 21, 51, 75, 200]
liststr_cmap = ["#79de79", "#a8e4ef", "#fcfc99", "#fb6962", "#ffffff"]

dict_heatmapopts = dict(color_levels=listfloat_color_levels, cmap=liststr_cmap)

heatmap = df.hvplot.heatmap(shared_axes=False).opts(
    title=str(listfloat_color_levels),
    clim=(0, 75),
    colorbar_opts={
        "major_label_overrides": {
            0: "low",
            11: "medium",
            21: "high",
            51: "higher",
            75: "like, even higher",
        },
        'ticker': FixedTicker(ticks=listfloat_color_levels)
    },
    **dict_heatmapopts,
)
hvdim_value = hv.Dimension("value", value_format=lambda x: "%.3f" % x)
label = hv.Labels(heatmap, vdims=hvdim_value).opts(
    padding=0, text_color="black", text_alpha=1.0, text_font_size="11pt"
)
listplots = []
listplots.append(heatmap * label)
# listplots.append(heatmap * label)
# gc.collect()
hv.Layout(listplots).cols(2)

image

2 Likes