Interactive Plot of DataFrame by index with Ipywidgets

Question

I have two DataFrames:

import numpy as np
import pandas as pd
import ipywidgets as widgets
import seaborn as sns

df = pd.DataFrame(np.random.rand(6,2), index=['Nr. 1', 'Nr. 2', 'Nr. 3', 'Nr. 4', 'Nr. 5', 'Nr. 6'], columns=['A', 'B', 'C', 'D'])

df2 = pd.DataFrame(np.array([[1, 2, 3, 4]]), columns=['a', 'b', 'c', 'd'])

I want to make an interactive plot where I can choose the row of the DataFrame by a dropdown of the index of df.

This is my approach:

Index= df.index.tolist()
Nr = widgets.Dropdown(options=Nr, value=Nr[0], description='Number:', disabled=False)
button = widgets.Button(description='Plot', disabled = False, button_style='', tooltip = 'Plotting', icon='check')

out = widgets.Output(layout={'border': '1px solid black'})
box = widgets.VBox([Nr, button])
def on_button_clicked(b):
    with out:
         ax = sns.regplot(x=df_2[0], y=df.loc[[Nr]])
button.on_click(on_button_clicked, False)
display(box)

However, I get this error: 'None of [Index([Dropdown(description='Nr:', index = 4, options=('Nr. 1', 'Nr. 2', 'Nr. 3', 'Nr. 4'), value = 'Nr. 1')], dtype = 'object', name='Nr')] are in the [index]'

What am I missing?

Answer 1

• seaborn.regplot works with long form data, as specified in the docs.
• A regression plot must have numeric values for the x and y axis
• See Jupyter Widgets for install instructions
• Referenced:
  • Convert categorical data in pandas dataframe
  • How to make an interactive bar plot with seaborn and ipywidgets
• Tested in python 3.10, pandas 1.4.2, matplotlib 3.5.1, seaborn 0.11.2, notebook 6.4.8, jupyterlab 3.3.2, jupyterlab_widgets 1.0.0, ipywidgets 7.6.4

import ipywidgets as widgets
from ipywidgets import interact  
import pandas as pd
import numpy as np
import seaborn as sns

# create wide dataframe
df = pd.DataFrame(np.random.rand(6, 4),
                  index=['Nr. 1', 'Nr. 2', 'Nr. 3', 'Nr. 4', 'Nr. 5', 'Nr. 6'],
                  columns=['A', 'B', 'C', 'D'])

# convert the dataframe to long form
df = df.reset_index().melt(id_vars='index')

# convert the categorical value to a number with cat.codes or factorize
df['fac'], cats = df.variable.factorize() 
# df['fac'] = df.variable.astype('category').cat.codes
# cats = df.variable.unique()

# display(df.head())
   index variable     value  fac
0  Nr. 1        A  0.700304    0
1  Nr. 2        A  0.375954    0
2  Nr. 3        A  0.168559    0
3  Nr. 4        A  0.962506    0
4  Nr. 5        A  0.503662    0

Interactive Plot

# get the unique values from the column used to select the data
idx = df['index'].unique()
@interact(Idx = idx)
def f(Idx):
    # select the relevant data
    data = df[df['index'].eq(Idx)]
    # plot
    ax = sns.regplot(data=data, x='fac', y='value')
    # set the x-ticks and labels
    ax.set_xticks(df.fac.unique(), cats)
    return ax