Python | numpy.nanmean() function

numpy.nanmean() function can be used to calculate the mean of array ignoring the NaN value. If array have NaN value and we can find out the mean without effect of NaN value.

Syntax: numpy.nanmean(a, axis=None, dtype=None, out=None, keepdims=))
Parameters:
a: [arr_like] input array 
axis: we can use axis=1 means row wise or axis=0 means column wise. 
out: output array 
dtype: data types of array 
overwrite_input: If True, then allow use of memory of input array a for calculations. The input array will be modified by the call to median. 
keepdims: If this is set to True, the axes which are reduced are left in the result as dimensions with size one. With this option, the result will broadcast correctly against the original a.
Returns: Returns the average of the array elements

Example #1: 

Python3

# Python code to demonstrate the
# use of numpy.nanmean
import numpy as np

# create 2d array with nan value.
arr = np.array([[20, 15, 37], [47, 13, np.nan]])

print("Shape of array is", arr.shape)

print("Mean of array without using nanmean function:",
                                           np.mean(arr))

print("Using nanmean function:", np.nanmean(arr))

Output:

Shape of array is (2, 3)
Mean of array without using nanmean function: nan
Using nanmean function: 26.4

Example #2: 

Python3

# Python code to demonstrate the
# use of numpy.nanmean
# with axis = 0
import numpy as np

# create 2d matrix with nan value
arr = np.array([[32, 20, 24],
                [47, 63, np.nan],
                [17, 28, np.nan],
                [10, 8, 9]])

print("Shape of array is", arr.shape)

print("Mean of array with axis = 0:",
             np.mean(arr, axis = 0))

print("Using nanmedian function:",
      np.nanmean(arr, axis = 0))

Output:

Shape of array is (4, 3)
Mean of array with axis = 0: [ 26.5   29.75    nan]
Using nanmedian function: [ 26.5   29.75  16.5 ]

Example #3: 

Python3

# Python code to demonstrate the
# use of numpy.nanmedian
# with axis = 1
import numpy as np

# create 2d matrix with nan value
arr = np.array([[32, 20, 24],
                [47, 63, np.nan],
                [17, 28, np.nan],
                [10, 8, 9]])

print("Shape of array is", arr.shape)

print("Mean of array with axis = 1:",
             np.mean(arr, axis = 1))

print("Using nanmedian function:",
      np.nanmean(arr, axis = 1))

Output:

Shape of array is (4, 3)
Mean of array with axis = 1: [ 25.33333333          nan          nan   9.        ]
Using nanmedian function: [ 25.33333333  55.          22.5          9.        ]