Customer Segmentation using K-Means Clustering

Data Source: Kaggle(Customer Loyality dataset)
Softwares: python- Numpy, Pandas, Scikitlearn


Outline

1. Understanding the objective

2. Requirements

3.Importing the neccessary libraries & data set

4. Data Cleaning and Exploring data Analysis

5.Model Training

6.Data Visualization


Understanding the Objective

K means clustering is an unsupervised machine learning algorithm that separates a given data into a specified number of clusters
in order to uncover hidden patterns in the data. This project attempts to divide the Mall customers into groups of similar characterisitics.
In this case, based on their spending score and Annual Income. In this case, it is a behavioural market segmentation and this is specifically useful in identifying which customer is more valuable to the business.


Uses of Clustering

  • Marketing: To help determiine who gets discounts, offers or promotion codes based on the groups.
  • Document Analysis: To group texts and documents according to similarity
  • Increase customer retention and decrease churn
  • Determine appropriate product pricing.
  • Real Estate: To understand and divide various property locations based on value and importance.

    • 2. Requirements

    Build a database to segment customers based on their Annual income and spending score.
  • How many categories of customers do we have?
    • The data includes the following features:
    1. Customer ID
    2. Customer Gender
    3. Customer Age
    4. Annual Income of the customer (in Thousand Dollars)
    5. Spending score (based on customer behaviour and spending nature)

    Importing the neccessary libraries and data set

    The cleaned dataset was imported in Excel csv format into the Google colaboratory notebook

    Data Cleaning and Exploring data Analysis

    A detailed Exploratory data analysis was done on the data using the Pandas profile Report. Overview variables Correlation
    Missing values: There are no missing values in the dataset Plot of Age with Total percentage of Income The plot shows that the age group on the legend as we scroll, each contributes the displayed percentage of the Total Income on the doughnut chart Specifying our data for x in terms of Annual Income and Spending score

    Model Training

    Choosing the number of clusters using the WCSS- Within clusters sum of squares
    Training the K-Means Clustering Model

    Data Visualization






    Summary of results

    From the results, we have identified with the colors, a total of 5 different clusters as formed from the model. The green segment is the customers with the least income and least spending score, in the same vein, the magenta segment is the customer category with the most income and most spending score.

    Social