Python Bokeh - interactive plot using select widget and interactive legend - empty html file

Question

I heave some issues plotting with Bokeh. There is a filter on 'category', thus plot should update on change. Also, the plot has an interactive legend.

There must be something wring with the code. I get an error "BAD-COLUMN_NAME" in jupyter notebook, but probably there are there issues too, and the html file is empty. Please, I need some help here. Thank you!

# Perform necessary imports
import pandas as pd
from bokeh.io import output_file, show
from bokeh.plotting import figure
from bokeh.models import CustomJS, HoverTool, ColumnDataSource, Select
from bokeh.palettes import Category20_20
from bokeh.layouts import layout, widgetbox, gridplot

#Dataframe
df = pd.DataFrame({'Reg': ['1', '2', '3', '4', '5', '6', '7', '8', '9', '10'],
        'Category': ['Cat1', 'Cat1', 'Cat1', 'Cat2', 'Cat2', 'Cat2', 'Cat3', 'Cat3', 'Cat3', 'Cat3'],
        'X': ['751', '673', '542', '762', '624', '536', '845', '626', '876', '233'],
        'Y': ['458', '316', '287', '303', '297', '564', '278', '834', '234', '623'],
        'Size': ['5', '9', '5', '8', '10', '22', '23', '12', '9', '20'],           
        'User': ['u1', 'u2', 'u3', 'u1', 'u2', 'u3', 'u1', 'u2', 'u3', 'u3']           
        })

# Make the ColumnDataSource
source = ColumnDataSource(data=dict(Reg=df['Reg'], Category=df['Category'], x=df['X'], y=df['Y'], Size=df['Size'], User=df['User']))
filteredSource = ColumnDataSource(data=dict(Reg=[], Category=[], x=[], y=[], Size=[], User=[]))

#Create widget
category = list(set(df['Category']))
category.sort()
category_select = Select(title="Category:", value=category[0], options=category)

#Callback
callback = CustomJS(args=dict(source=source, filteredSource=filteredSource, 
                              category_select=category_select), code="""
    const data = source.data;
    var f = cb_obj.value;
    const df2 = filteredSource.data;
    df2['category']=[]

    for(i = 0; i < data['category'].length;i++){

    if(data['category'][i]==f){
        df2['category'].push(data['category'][i])
    }
    }

    filteredSource.change.emit()

""")

category_select.js_on_change('value', callback)

# Create the figure: p1
p1 = figure(x_axis_label='x)', y_axis_label='y', 
            plot_width=450, plot_height=450, tools=['box_select', HoverTool(tooltips='Size: @size')])
# Add a circle glyph to p1

users=list(set(df['User']))
for name, color in zip(users, Category20_20):
    user_df = df[df['User'] == name]
    p1.circle(x='X', y='Y', size='Size',
             color=color, alpha=0.8, source=filteredSource, legend=name)

p1.legend.location = "top_right"
p1.legend.click_policy="hide"

#layout
layout = gridplot([widgetbox(category_select), p1], ncols=2, sizing_mode='scale_both')
show(layout)

Eugene Pakhomov · Accepted Answer · 2020-05-26 19:08:52Z

1

You have inconsistent character cases in column names. You use e.g. x and X. You have to stick to one variant and use it everywhere - in data sources, in column specifications when creating glyph renderers, in CustomJS code, etc.

Also, your CustomJS code is wrong. i = 0 will result in an error - you should use let i = 0 instead. And you also update just the Category column whereas you have to update all of the columns to make sure that they're of the same length.

As a side note, judging by this and the other question, you may find this documentation section rather useful: https://docs.bokeh.org/en/latest/docs/user_guide/data.html#filtering-data. The approach described there helps prevent writing JS code and unneded data churning. There are multiple examples here and on Bokeh's Discourse.

answered May 26, 2020 at 19:08

Eugene Pakhomov

11.2k3 gold badges32 silver badges57 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

Anamaria Over a year ago

Thank you for your feedback, Eugene!

Collectives™ on Stack Overflow

Python Bokeh - interactive plot using select widget and interactive legend - empty html file

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related