Create Correlation Matrix in Python by Traversing Each Line

A correlation matrix is a table containing correlation coefficients for many variables. Each cell in the table represents the correlation between two variables. The value might range between -1 and 1. A correlation matrix is used for summarizing the data, diagnose the advanced analysis, and as an input for a more complicated study.

Correlation matrix is used to represent the relationship between the variables in the data set. It is a type of matrix that helps programmers analyze the relationship between data components. It represents the correlation coefficient between 0 and 1.

A positive value implies a good correlation, a negative value shows a weak/low correlation, and a value of zero(0) indicates no dependency between the given set of variables.

The Regression Analysis and Correlation Matrix showed the following observations ?

• Recognize the relationship between the independent variables in the data set.

• Helps in the selection of significant and non-redundant variables from a data set.

• This only applies to variables that are numeric or continuous.

In this article, we will show you how to create a correlation matrix using python.

Assume we have taken a CSV file with the name starbucksMenu.csv consisting of some random data. We need to create a correlation matrix for the specified columns in a dataset and plot the correlation matrix.

Input File Data

starbucksMenu.csv

Creating a Correlation Matrix

We will plot the correlation matrix for the three columns of the dataset which are independent continuous variables.

• Carb
• Protein
• Sodium

Algorithm (Steps)

Following are the Algorithm/steps to be followed to perform the desired task ?

• Importing the os, pandas, NumPy, and seaborn libraries.

• Read the given CSV file using the read_csv() function(loads a CSV file as a pandas dataframe).

• Create the list of columns from the given dataset for which the correlation matrix must be created.

• Create a correlation matrix using the corr() function(It calculates the pairwise correlation of all columns in a data frame. Any na(null) values are automatically filtered out. It is discarded for any non-numeric data type columns in the dataframe).

• Print the correlation matrix of the specified columns of the dataset.

• Plot the correlation matrix using the heatmap() function(For each value to be plotted, a heatmap has values indicating several shades of the same color. The darker colors of the chart typically represent higher values than the lighter shades. A completely different color can likewise be utilized for a significantly different value) of the seaborn library.

Importing the Dataset into a Pandas Dataframe

We are now first importing any sample dataset(here we are using starbucksMenu.csv ) into pandas dataframe and printing it.

Example 1

# Import pandas module as pd using the import keyword
import pandas as pd
# Reading a dataset
givenDataset = pd.read_csv('starbucksMenu.csv')
print(givenDataset)

Output

Creating correlation matrix after importing the dataset

The following program finds out how to create a correlation matrix for the given dataset, prints them, and plots the correlation matrix ?

Example 2

import os
# Importing pandas module
import pandas as pd
import numpy as np
import seaborn 
# Reading a dataset
givenDataset = pd.read_csv('starbucksMenu.csv')
# Assigning the list of columns from the dataset 
numericColumns = ['Carb','Protein','Sodium']

# Creating a correlation matrix 
correlationMatrix  = givenDataset.loc[:,numericColumns].corr()
# Printing the correlation matrix.
print(correlationMatrix)
# Displaying the correlation matrix
seaborn.heatmap(correlationMatrix, annot=True)

Output

On executing, the above program will generate the following output ?

You learned how to compute a correlation matrix using Python and Pandas in this tutorial. Along with that you have learned how to generate a correlation matrix using the Pandas corr() method and also how to utilize the Seaborn library's heatmap function to show a matrix, allowing you to better visualize and understand the data at a glance.