Interactive Bokeh Legends

Source

Introduction

This is a reproduction of the bokeh interactive legends example.

Set Up

Imports

Python

I have a class that I use to embed the bokeh and I'm going to make it easier to reuse by binding some values to it with partial.

from functools import partial

PyPi

  • Bokeh

    The Spectral color scheme from bokeh.palettes is a categorical color scheme from Color Brewer, a color scheme helper for cartographers.

    from bokeh.palettes import Spectral4
from bokeh.plotting import figure
from bokeh.sampledata.stocks import AAPL, IBM, MSFT, GOOG
  • Pandas And HoloViews

    Pandas is for the data, holoviews is because I created the EmbedBokeh class expecting to always use it..

    import holoviews
import pandas

My Stuff

This is a convenience class for embedding the javascript that bokeh creates into this post.

from bartleby_the_penguin.tangles.embed_bokeh import EmbedBokeh

The Embedder

The files_path is where the javascript needs to be stored for nikola to copy it to the right place.

files_path = Path("../../files/posts/libraries/interactive-bokeh-legends/")
Embed = partial(
    EmbedBokeh,
    folder_path=files_path)

The Bokeh Backend

For some reason the backend is defaulting to matplotlib so this fixes it.

holoviews.extension("bokeh")

Hide Plots

This creates a plot where the entry fore each line in the legend becomes a button that toggles the plot's visibility.

plot = figure(plot_width=800, plot_height=400, 
              x_axis_type="datetime", 
              tools="hover,pan,wheel_zoom,box_zoom,reset")
plot.title.text = "Some Stock Prices (Click On the Legend To Hide Plots)"

for data, name, color in zip([AAPL, IBM, MSFT, GOOG], "Apple IBM Microsoft Google".split(), Spectral4):
    frame = pandas.DataFrame(data)
    frame["date"] = pandas.to_datetime(frame["date"])
    plot.line(frame["date"], frame["close"], line_width=2, color=color, alpha=0.8, legend=name)
plot.legend.location = "top_left"
plot.legend.click_policy = "hide"
plot.title.text_font_size = "14pt"

Output The Plot

Fade Plots

Clicking on the line in the legend will change the alpha-value for the line (to make it less visible). This has the effect of keeping the hovertool working for the line, while hiding it disables the hover tool.

fade_plot = figure(plot_width=800, plot_height=400, 
                   x_axis_type="datetime", 
                   tools="hover,pan,wheel_zoom,box_zoom,reset")
fade_plot.title.text = "Some Stock Prices (Click On the Legend To Mute Plots)"
fade_plot.title.text_font_size = "14pt"

for data, name, color in zip([AAPL, IBM, MSFT, GOOG], "Apple IBM Microsoft Google".split(), Spectral4):
    frame = pandas.DataFrame(data)
    frame["date"] = pandas.to_datetime(frame["date"])
    fade_plot.line(frame["date"], frame["close"], line_width=2, color=color, muted_alpha=0.2, muted_color=color, alpha=0.8, legend=name)
fade_plot.legend.location = "top_left"
fade_plot.legend.click_policy = "mute"

Output The Plot