Correlation with corr() in Data Analysis Python

Introduction

Correlation helps us see how two things change together. It tells if one thing goes up when the other goes up or down.

To check if study time and test scores are related.

To find if temperature and ice cream sales move together.

To see if height and weight have a connection.

To understand if advertising budget affects sales.

To explore relationships between different features in a dataset before modeling.

Syntax

Data Analysis Python

DataFrame.corr(method='pearson')

This method calculates correlation between numeric columns in a DataFrame.

The default method is 'pearson', which measures linear correlation.

Examples

Calculate Pearson correlation between all numeric columns in the DataFrame df.

Data Analysis Python

df.corr()

Explicitly specify Pearson correlation (default).

Data Analysis Python

df.corr(method='pearson')

Calculate Kendall Tau correlation, useful for small datasets or non-linear relationships.

Data Analysis Python

df.corr(method='kendall')

Calculate Spearman rank correlation, which measures monotonic relationships.

Data Analysis Python

df.corr(method='spearman')

Sample Program

This code creates a small dataset with study hours, test scores, and sleep hours. Then it calculates the correlation matrix to see how these columns relate.

Data Analysis Python

import pandas as pd

# Create a simple dataset
data = {
    'study_hours': [2, 3, 5, 8, 10],
    'test_score': [50, 55, 65, 80, 90],
    'sleep_hours': [7, 6, 5, 6, 7]
}
df = pd.DataFrame(data)

# Calculate correlation matrix
corr_matrix = df.corr()
print(corr_matrix)

Important Notes

Correlation values range from -1 to 1.

A value close to 1 means strong positive relation, close to -1 means strong negative relation, and around 0 means no relation.

Only numeric columns are used in correlation calculation.

Summary

Use corr() to find relationships between numeric columns.

Correlation helps understand how variables move together.

Values near 1 or -1 show strong relationships; near 0 means weak or no relationship.