Search code examples
pythonpandasbokeh

I want to add two y axis values in a bokeh line plot

whats wrong with this line of code..please help

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

output_file ('newfile.html')

data = pd.read_excel(r'C:\Users\ASyed\OneDrive -      NKT\PythonProject\bokehPractise\newfile\19AJ100429-GC3-FR-003-A1-KP240.000-KP248.831-SL-AT.xlsx', \
sheet_name= 'Listing')
df = pd.DataFrame(data)
df.columns = [x.replace("\n", " ") for x in df.columns.to_list()]

SOURCE = ColumnDataSource(data = df)
p = figure (plot_width = 800, plot_height = 600)


p.line(x= 'KP [km]', y =[['DOL [m]'], ['DOC [m]']], source = SOURCE )
p.title.text = 'DOL Visualization'
p.xaxis.axis_label = 'Kilometer Point'
p.yaxis.axis_label = 'DOC'

show(p)

I am using pandas and bokeh. i want a plot with two y axis lines derived from two different columns. Above code gives following error and a plank bokeh plot in html.

Expected y to reference fields in the supplied data source.

When a 'source' argument is passed to a glyph method, values that are sequences
(like lists or arrays) must come from references to data columns in the source.

For instance, as an example:

    source = ColumnDataSource(data=dict(x=a_list, y=an_array))

    p.circle(x='x', y='y', source=source, ...) # pass column names and a source

Alternatively, *all* data sequences may be provided as literals as long as a
source is *not* provided:

    p.circle(x=a_list, y=an_array, ...)  # pass actual sequences and no source
Solution

  • When you are using source as an input to your figure, you are only allowed to add strings as a pointer to the data.

    In your case use

    p.line(x= 'KP [km]', y ='DOL [m]', source = SOURCE )

    instead of

    p.line(x= 'KP [km]', y =[['DOL [m]'], ['DOC [m]']], source = SOURCE )

    If you want to add multiple lines to one figure, call the p.line() once for each line.

    Minimal Example

    import pandas as pd

from bokeh.plotting import figure, show, output_notebook
from bokeh.models import ColumnDataSource
output_notebook()

df = pd.DataFrame({'a':[1,2,3], 'b':[4,5,6]})
source = ColumnDataSource(df)

p = figure(width=300, height=300)
p.line(x='index', y='a', source=source)
p.line(x='index', y='b', source=source, color='red')

show(p)

    multiple line plot

    See also

    Check out the multiline example on the official documentation.