CODEWITHSUNDEEP
pythonnumpyseaborn
Stephen Williams
Stephen Williams

Reputation: 621

Seaborn Grouped Violin Plot WITHOUT pandas

For reasons I won't get into I need to make violin plots without using a pandas dataframe. For example I have the following ndarray and categories.

import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt


data = np.random.randn(5, 3)
category = np.array(["yes", "no", "no", "no", "yes", "yes","yes", "no", "yes", "yes", "yes", "no", "no", "no", "no"])

ax = sns.violinplot(data = data)
plt.show()

Results in an ungrouped violin plot.

violin plots

However, I'd like to use the categorical data to make a grouped violin plot

ax = sns.violinplot(data = data, x = category)
plt.show()

Gives an error AttributeError: 'numpy.ndarray' object has no attribute 'get'. Is there any way around this without pandas?

Upvotes: 2

Views: 1385

Answers (1)

David Erickson
David Erickson

Reputation: 16683

  1. Do not use the data parameter if using multiple numpy arrays for x, y and hue.
  2. From y, you can create an array of indices with np.nonzero.
  3. Make sure all of your np.arrays are one-dimensional with .flatten(). For example I flatten your array of random floats from a shape of 5,3 to 15,1; Otherwise, you will get an error since the arrays have different shapes and Seaborn doesn't have a way to figure it out as it can with a pandas dataframe.

Likewise, if you pass three (5,3) arrays to x, y and hue, then Seaborn won't know what to do. So, you must either a) FLATTEN all arrays and make them equal length of (15,0) OR b) use a pandas dataframe.

import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt


y = np.random.randn(5, 3)
x = np.nonzero(y)[-1]
y = y.flatten()
hue = np.array(["yes", "no", "no", "no", "yes", "yes","yes", "no", "yes", "yes", "yes", "no", "no", "no", "no"])
sns.violinplot(x=x, y=y, hue=hue)

enter image description here

print(x,'\n\n',y,'\n\n',hue)

[0 1 2 0 1 2 0 1 2 0 1 2 0 1 2] 

 [-0.28618123 -1.18132595  0.70535902  0.90685532 -1.27258432  0.90417094
  3.03506025  0.99796779  0.20247628  0.43226169  0.25005372 -0.9923336
 -0.43102785 -0.17117549 -0.16147393] 

 ['yes' 'no' 'no' 'no' 'yes' 'yes' 'yes' 'no' 'yes' 'yes' 'yes' 'no' 'no'
 'no' 'no']

Upvotes: 3

Related Questions