How to Calculate Correlation Using NumPy in Python

You can calculate correlation in NumPy using the numpy.corrcoef() function, which returns the correlation matrix of input arrays. Pass your data arrays to this function to get the Pearson correlation coefficients between them.

Syntax

The basic syntax to calculate correlation using NumPy is:

• numpy.corrcoef(x, y=None, rowvar=True, bias=deprecated, ddof=deprecated, dtype=None)

Here:

• x: Input array or sequence of variables.
• y: Optional second array to compute correlation with x.
• rowvar: If True (default), each row represents a variable, and columns are observations.

The function returns a correlation matrix showing Pearson correlation coefficients between variables.

python

import numpy as np

# Syntax example
# corrcoef returns correlation matrix
corr_matrix = np.corrcoef(x, y=None, rowvar=True)

Example

This example shows how to calculate the correlation between two data arrays using numpy.corrcoef(). It prints the correlation matrix and the correlation coefficient between the two arrays.

python

import numpy as np

# Two sample data arrays
x = np.array([1, 2, 3, 4, 5])
y = np.array([5, 4, 3, 2, 1])

# Calculate correlation matrix
corr_matrix = np.corrcoef(x, y)

print("Correlation matrix:\n", corr_matrix)

# Extract correlation coefficient between x and y
corr_xy = corr_matrix[0, 1]
print(f"Correlation coefficient between x and y: {corr_xy:.2f}")

Output

Correlation matrix:
 [[ 1. -1.]
 [-1.  1.]]
Correlation coefficient between x and y: -1.00

Common Pitfalls

Common mistakes when calculating correlation with NumPy include:

• Passing 1D arrays without specifying y when you want correlation between two variables.
• Confusing rowvar parameter when working with 2D arrays.
• Expecting a single number instead of a matrix when passing multiple variables.

Always check the shape of your input arrays and understand that corrcoef returns a matrix, not a single value unless you extract it.

python

import numpy as np

# Wrong: Passing only one array expecting single correlation value
x = np.array([1, 2, 3, 4, 5])
# This returns 1 because it's correlation of x with itself
print(np.corrcoef(x))

# Right: Pass two arrays to get correlation between them
y = np.array([5, 4, 3, 2, 1])
print(np.corrcoef(x, y))

Output

[[1.]]
[[ 1. -1.]
 [-1.  1.]]

Quick Reference

Summary tips for using numpy.corrcoef():

• Use np.corrcoef(x, y) to get correlation matrix between two arrays.
• For multiple variables, pass a 2D array with variables as rows (default) or columns (set rowvar=False).
• Extract correlation coefficient from the matrix by indexing.
• Correlation values range from -1 (perfect negative) to 1 (perfect positive).

Key Takeaways

Use numpy.corrcoef() to calculate Pearson correlation coefficients between arrays.

Pass two arrays to get their correlation matrix; extract the value you need from the matrix.

Remember correlation values range from -1 to 1 indicating strength and direction.

Check array shapes and the rowvar parameter to avoid confusion with multi-variable data.

corrcoef returns a matrix, not a single number unless you extract the specific coefficient.