How to fix 'numpy.ndarray' object has no attribute 'get_figure' when plotting subplots

Question

I have written the following code to plot 6 pie charts in different subplots, but I get an error. This code works correctly if I use it to plot only 2 charts, but produces an an error for anything more than that.

I have 6 categorical variables in my dataset, the names of which are stored in the list cat_cols. The charts are to be plotted from the training data train.

CODE

fig, axes = plt.subplots(2, 3, figsize=(24, 10))

for i, c in enumerate(cat_cols):
  
  train[c].value_counts()[::-1].plot(kind = 'pie', ax=axes[i], title=c, autopct='%.0f', fontsize=18)
  axes[i].set_ylabel('')
    
plt.tight_layout()

ERROR

AttributeError: 'numpy.ndarray' object has no attribute 'get_figure'

How do we rectify this?

Answer 1

• The issue is plt.subplots(2, 3, figsize=(24, 10)) creates two groups of 3 subplots, not one group of six subplots.

array([[<AxesSubplot:xlabel='radians'>, <AxesSubplot:xlabel='radians'>, <AxesSubplot:xlabel='radians'>],
       [<AxesSubplot:xlabel='radians'>, <AxesSubplot:xlabel='radians'>, <AxesSubplot:xlabel='radians'>]], dtype=object)

• Unpack all of the subplot arrays from axes, using axes.ravel().
  • numpy.ravel, which returns a flattened array.
  • A list comprehension will also work, axe = [sub for x in axes for sub in x]
  • In practical terms, axes.ravel(), axes.flat, and axes.flatten(), can be used similarly. See What is the difference between flatten and ravel functions in numpy? & numpy difference between flat and ravel().
• Assign each plot to one of the subplots in axe.
• How to resolve AttributeError: 'numpy.ndarray' object has no attribute 'get_figure' when plotting subplots is a similar issue.

import pandas as pd
import numpy as np

# sinusoidal sample data
sample_length = range(1, 6+1)
rads = np.arange(0, 2*np.pi, 0.01)
data = np.array([np.sin(t*rads) for t in sample_length])
df = pd.DataFrame(data.T, index=pd.Series(rads.tolist(), name='radians'), columns=[f'freq: {i}x' for i in sample_length])

# crate the figure and axes
fig, axes = plt.subplots(2, 3, figsize=(24, 10))

# unpack all the axes subplots
axe = axes.ravel()

# assign the plot to each subplot in axe
for i, c in enumerate(df.columns):
    df[c].plot(ax=axe[i])

Answer 2

As was mentioned, you get two groups of 3 plots with plt.subplots. To create a one dimensional array of axes, you can also use numpy.flatten method:

fig, axes = plt.subplots(2, 3, figsize=(24, 10))

axe = axes.flatten()
for i, c in enumerate(cat_cols):

    train[c].value_counts()[::-1].plot(kind = 'pie', ax=axe[i], title=c, 
                                  autopct='%.0f', fontsize=18)
    axe[i].set_ylabel('')

plt.tight_layout()