
- NumPy - Home
- NumPy - Introduction
- NumPy - Environment
- NumPy Arrays
- NumPy - Ndarray Object
- NumPy - Data Types
- NumPy Creating and Manipulating Arrays
- NumPy - Array Creation Routines
- NumPy - Array Manipulation
- NumPy - Array from Existing Data
- NumPy - Array From Numerical Ranges
- NumPy - Iterating Over Array
- NumPy - Reshaping Arrays
- NumPy - Concatenating Arrays
- NumPy - Stacking Arrays
- NumPy - Splitting Arrays
- NumPy - Flattening Arrays
- NumPy - Transposing Arrays
- NumPy Indexing & Slicing
- NumPy - Indexing & Slicing
- NumPy - Indexing
- NumPy - Slicing
- NumPy - Advanced Indexing
- NumPy - Fancy Indexing
- NumPy - Field Access
- NumPy - Slicing with Boolean Arrays
- NumPy Array Attributes & Operations
- NumPy - Array Attributes
- NumPy - Array Shape
- NumPy - Array Size
- NumPy - Array Strides
- NumPy - Array Itemsize
- NumPy - Broadcasting
- NumPy - Arithmetic Operations
- NumPy - Array Addition
- NumPy - Array Subtraction
- NumPy - Array Multiplication
- NumPy - Array Division
- NumPy Advanced Array Operations
- NumPy - Swapping Axes of Arrays
- NumPy - Byte Swapping
- NumPy - Copies & Views
- NumPy - Element-wise Array Comparisons
- NumPy - Filtering Arrays
- NumPy - Joining Arrays
- NumPy - Sort, Search & Counting Functions
- NumPy - Searching Arrays
- NumPy - Union of Arrays
- NumPy - Finding Unique Rows
- NumPy - Creating Datetime Arrays
- NumPy - Binary Operators
- NumPy - String Functions
- NumPy - Matrix Library
- NumPy - Linear Algebra
- NumPy - Matplotlib
- NumPy - Histogram Using Matplotlib
- NumPy Sorting and Advanced Manipulation
- NumPy - Sorting Arrays
- NumPy - Sorting along an axis
- NumPy - Sorting with Fancy Indexing
- NumPy - Structured Arrays
- NumPy - Creating Structured Arrays
- NumPy - Manipulating Structured Arrays
- NumPy - Record Arrays
- Numpy - Loading Arrays
- Numpy - Saving Arrays
- NumPy - Append Values to an Array
- NumPy - Swap Columns of Array
- NumPy - Insert Axes to an Array
- NumPy Handling Missing Data
- NumPy - Handling Missing Data
- NumPy - Identifying Missing Values
- NumPy - Removing Missing Data
- NumPy - Imputing Missing Data
- NumPy Performance Optimization
- NumPy - Performance Optimization with Arrays
- NumPy - Vectorization with Arrays
- NumPy - Memory Layout of Arrays
- Numpy Linear Algebra
- NumPy - Linear Algebra
- NumPy - Matrix Library
- NumPy - Matrix Addition
- NumPy - Matrix Subtraction
- NumPy - Matrix Multiplication
- NumPy - Element-wise Matrix Operations
- NumPy - Dot Product
- NumPy - Matrix Inversion
- NumPy - Determinant Calculation
- NumPy - Eigenvalues
- NumPy - Eigenvectors
- NumPy - Singular Value Decomposition
- NumPy - Solving Linear Equations
- NumPy - Matrix Norms
- NumPy Element-wise Matrix Operations
- NumPy - Sum
- NumPy - Mean
- NumPy - Median
- NumPy - Min
- NumPy - Max
- NumPy Set Operations
- NumPy - Unique Elements
- NumPy - Intersection
- NumPy - Union
- NumPy - Difference
- NumPy Random Number Generation
- NumPy - Random Generator
- NumPy - Permutations & Shuffling
- NumPy - Uniform distribution
- NumPy - Normal distribution
- NumPy - Binomial distribution
- NumPy - Poisson distribution
- NumPy - Exponential distribution
- NumPy - Rayleigh Distribution
- NumPy - Logistic Distribution
- NumPy - Pareto Distribution
- NumPy - Visualize Distributions With Sea born
- NumPy - Matplotlib
- NumPy - Multinomial Distribution
- NumPy - Chi Square Distribution
- NumPy - Zipf Distribution
- NumPy File Input & Output
- NumPy - I/O with NumPy
- NumPy - Reading Data from Files
- NumPy - Writing Data to Files
- NumPy - File Formats Supported
- NumPy Mathematical Functions
- NumPy - Mathematical Functions
- NumPy - Trigonometric functions
- NumPy - Exponential Functions
- NumPy - Logarithmic Functions
- NumPy - Hyperbolic functions
- NumPy - Rounding functions
- NumPy Fourier Transforms
- NumPy - Discrete Fourier Transform (DFT)
- NumPy - Fast Fourier Transform (FFT)
- NumPy - Inverse Fourier Transform
- NumPy - Fourier Series and Transforms
- NumPy - Signal Processing Applications
- NumPy - Convolution
- NumPy Polynomials
- NumPy - Polynomial Representation
- NumPy - Polynomial Operations
- NumPy - Finding Roots of Polynomials
- NumPy - Evaluating Polynomials
- NumPy Statistics
- NumPy - Statistical Functions
- NumPy - Descriptive Statistics
- NumPy Datetime
- NumPy - Basics of Date and Time
- NumPy - Representing Date & Time
- NumPy - Date & Time Arithmetic
- NumPy - Indexing with Datetime
- NumPy - Time Zone Handling
- NumPy - Time Series Analysis
- NumPy - Working with Time Deltas
- NumPy - Handling Leap Seconds
- NumPy - Vectorized Operations with Datetimes
- NumPy ufunc
- NumPy - ufunc Introduction
- NumPy - Creating Universal Functions (ufunc)
- NumPy - Arithmetic Universal Function (ufunc)
- NumPy - Rounding Decimal ufunc
- NumPy - Logarithmic Universal Function (ufunc)
- NumPy - Summation Universal Function (ufunc)
- NumPy - Product Universal Function (ufunc)
- NumPy - Difference Universal Function (ufunc)
- NumPy - Finding LCM with ufunc
- NumPy - ufunc Finding GCD
- NumPy - ufunc Trigonometric
- NumPy - Hyperbolic ufunc
- NumPy - Set Operations ufunc
- NumPy Useful Resources
- NumPy - Quick Guide
- NumPy - Cheatsheet
- NumPy - Useful Resources
- NumPy - Discussion
- NumPy Compiler
NumPy nanmean() Function
The NumPy nanmean() function computes the arithmetic mean (average) of the elements in an array along a specified axis, ignoring NaN (Not a Number) values. The mean is the sum of the data values divided by the number of valid (non-NaN) data points. By default, the function computes the mean over the flattened array, or over the specified axis if provided.
If all elements in the specified slice are NaN, the function returns NaN and raises a RuntimeWarning. When the array contains integers, the function converts these integers into a 64-bit floating-point format (float64) before performing the calculations. This ensures higher precision in the computation of the average. After the calculation, the returned result is also in the float64 format, even if the original array was composed of integers.
In statistics, the mean (also known as the average) is the sum of the valid data values divided by the number of valid data points. The formula is mean = (sum of all valid elements) / (number of valid elements).
For a one-dimensional array, the mean is computed over all non-NaN elements. For multi-dimensional arrays, the mean is computed along the specified axis while ignoring NaN values.
Syntax
Following is the syntax of the NumPy nanmean() function −
numpy.nanmean(a, axis=None, dtype=None, out=None, keepdims=<no value>,where=<no value>)
Parameters
Following are the parameters of the NumPy nanmean() function −
- a: Input array, which can be a NumPy array, list, or a scalar value. NaN values are ignored.
- axis (optional): Axis along which to compute the mean. Default is None, which means the mean is computed over the entire array.
- dtype (optional): The data type used to compute the mean. Default is None, meaning the data type of the input array is used.
- out (optional): A location into which the result is stored. If provided, it must have the same shape as the expected output.
- keepdims (optional): If True, the reduced dimensions are retained as dimensions of size one in the output. Default is False.
- where (optional): A boolean array specifying the elements to include in the calculation.
Return Values
This function returns the mean of the input array, ignoring NaN values. If the axis parameter is specified, the mean is computed along that axis. The result is a scalar for one-dimensional input and an array for multi-dimensional input.
Example
Following is a basic example to compute the mean of an array while ignoring NaN values using the NumPy nanmean() function −
import numpy as np # input array with NaN values x = np.array([1, 2, np.nan, 4, 5]) # applying nanmean result = np.nanmean(x) print("Mean Result (ignoring NaN):", result)
Output
Following is the output of the above code −
Mean Result (ignoring NaN): 3.0
Example: Specifying an Axis
The nanmean() function can compute the mean along a specific axis of a multi-dimensional array, ignoring NaN values. In the following example, we compute the mean along axis 0 (columns) and axis 1 (rows) of a 2D array −
import numpy as np # 2D array with NaN values x = np.array([[1, 2, np.nan], [4, np.nan, 6], [7, 8, 9]]) # applying nanmean along axis 0 (columns) result_axis0 = np.nanmean(x, axis=0) # applying nanmean along axis 1 (rows) result_axis1 = np.nanmean(x, axis=1) print("Mean along axis 0 (ignoring NaN):", result_axis0) print("Mean along axis 1 (ignoring NaN):", result_axis1)
Output
Following is the output of the above code −
Mean along axis 0 (ignoring NaN): [4. 5. 7.5] Mean along axis 1 (ignoring NaN): [1.5 5. 8. ]
Example: Usage of 'keepdims' Parameter
The keepdims parameter allows the result to retain the reduced dimensions as size one. This is useful for broadcasting the result back to the original shape. In the following example, we have computed the mean along axis 0 and retain the reduced dimensions −
import numpy as np # 2D array with NaN values x = np.array([[1, 2, np.nan], [4, np.nan, 6], [7, 8, 9]]) result = np.nanmean(x, axis=0, keepdims=True) print("Mean with keepdims=True (ignoring NaN):", result)
Output
Following is the output of the above code −
Mean with keepdims=True (ignoring NaN): [[4. 5. 7.5]]
Example: Plotting 'nanmean()' Function
In the following example, we have plotte the behavior of the nanmean() function, calculating and visualizing the mean for different input arrays with some NaN values using numpy and matplotlib.pyplot −
import numpy as np import matplotlib.pyplot as plt # Input array with some NaN values x = np.linspace(0, 10, 100) y = np.full_like(x, np.nanmean(x)) plt.plot(x, y, label="Mean (ignoring NaN)") plt.title("Nanmean Function") plt.xlabel("Input") plt.ylabel("Mean Value") plt.legend() plt.grid() plt.show()
Output
The plot demonstrates the constant nature of the mean value across the input range, while ignoring NaN values −
