Search code examples
pythonbokeh

How to Make Event Plot using Python Bokeh Library?

I am trying to make event plot using Bokeh Library. I know it is possible using matplotlib.pyplot.eventplot, But I need more interactivity features that is available in Bokeh

I came across the screenshot of such a chart in this Website. So it must be possible in Bokeh. Bokeh Event Plot

How do I make a similar plot using Bokeh Library?

Solution

  • The idea is very basic. You can use the rect() function of the figure object, where you have to define x, y and width. In your case the figure should have an x-axis in datetime-format. The width is your duration of on task.

    Minimal example

    import pandas as pd
from bokeh.plotting import figure, show
from bokeh.models import ColumnDataSource

df = pd.DataFrame({'y':[1, 2, 3, 4], 
                   'start':[0, 1, 1, 10], 
                   'duration':[1, 2, 1, 5], 
                   'color':['green', 'red', 'blue', 'red']
                  })
>>>df
   y  start  duration  color
0  1      0         1  green
1  2      1         2    red
2  3      1         1   blue
3  4     10         5    red

p = figure(
    width=350,
    height=350,
    title="Task Stream",
    toolbar_location="above",
    x_axis_type="datetime",
)

p.rect(
    source=ColumnDataSource(df),
    x="start",
    y="y",
    width="duration",
    height=0.4,
    fill_color="color",
    line_color="color",
)
show(p)

    Output

    Basic example for streaming data

    Comment

    As I mentionend in the comment, the figure you have shared is generated using the dask package which internally uses bokeh. They use the same idea I described above, but of course the dask project offers much more than this on figure and maby solves more than one problem you have.