Search code examples
pythonbokehholoviews

step plot using holoviews bokeh

I'm trying to add two step plots to the same Bokeh plot. Each step plot has different x values though, so the Step function in the bkcharts package doesnt work. I wanted to overlay two single Step charts produced by the bkcharts package, but it seems that you can't create a step chart with holoviews. Is there an easy workaround?

import bokeh.plotting as bk
from bokeh.charts import Step
import pandas as pd

data1 = pd.DataFrame({'time': [1,3,5,7], 'value': [3,4,1,3]})
data2 = pd.DataFrame({'time': [2,4,6,9], 'value': [2,1,4,2]})


step1 = Step(data1, x='time', y='value')
step2 = Step(data2, x='time', y='value')

bk.show(Step(data1, x='time', y='value'))
bk.show(Step(data2, x='time', y='value'))

I would like to do something like step1 * step2

Solution

  • HoloViews does let you do a step chart, using options to the Curve element. See the example in the Gallery:

    %%opts Curve [width=600] NdOverlay [legend_position='right']
hv.NdOverlay({interp: hv.Curve(points[::8])(plot=dict(interpolation=interp))
              for interp in ['linear', 'steps-mid', 'steps-pre', 'steps-post']})

    enter image description here

    You can do the same with a regular Overlay (the * operator) too; the key is to set the interpolation option to one of the available types of steps.