Customer Value Segmentation with RFM and K-Means Clustering

Project Overview

This project implements customer segmentation using RFM (Recency, Frequency, Monetary) analysis combined with K-Means clustering. The goal is to categorize customers into meaningful segments, such as high value, low value, and potentially loyal customers, by leveraging their transactional data. These insights help businesses target specific customer segments with personalized marketing strategies.

Dataset

The dataset used for this analysis includes customer transactions from an online retail store. The primary features include:

• Invoice: Invoice number of the transaction.
• StockCode: Unique identifier of the product.
• Description: Product description.
• Quantity: Number of units purchased.
• InvoiceDate: The date and time of the transaction.
• UnitPrice: Price per product unit.
• Customer ID: Unique identifier for each customer.
• Country: Customer's country of origin.

Data Cleaning

• Outliers: Outliers were identified and removed from the dataset to ensure accurate segmentation results.
• Missing Values: Rows with missing Customer ID were handled before analysis.

Steps Involved

1. Data Preprocessing:

   • Removal of outliers to ensure that the analysis is not skewed by extreme values.
   • Cleaning the dataset by filtering out incomplete records.

2. RFM Calculation:

   • Recency: How long it has been since a customer's last purchase.
   • Frequency: The total number of transactions a customer has made.
   • Monetary: The total amount a customer has spent.

3. Revenue Metrics Calculation: In addition to RFM, several key revenue metrics are calculated for each customer:

   • LastMonthRevenue: Revenue generated by the customer in the last month.
   • ThreeMonthsRevenue: Revenue generated in the last three months.
   • SixMonthsRevenue: Revenue generated in the last six months.
   • TwelveMonthsRevenue: Revenue generated over the last twelve months.
   • LifeTimeRevenue: Total revenue from the customer's lifetime of purchases.

4. K-Means Clustering:

   • K-Means clustering is applied to segment the customers into 3 distinct clusters based on RFM and revenue metrics.
   • The optimal number of clusters is determined using techniques like the elbow method.

5. Customer Segmentation:

   • After clustering, the customers are grouped into three segments:
     • High Value: Customers with high revenue and frequent purchases.
     • Low Value: Customers with low engagement or spend.
     • Potentially Loyal: Customers who show signs of loyalty but aren't high spenders yet.

6. Visualization:

   • The customer segments are visualized using plots and charts to better understand their distribution and characteristics.
   • Revenue trends and customer behaviors are represented graphically.

Results

The segmentation using K-Means provided meaningful insights into customer behavior, allowing for targeted strategies aimed at improving customer retention and boosting revenue.

How to Run the Code

1. Clone the repository:

   git clone <repo_url>
   cd <repo_folder>

2. Install the required libraries:

   pip install -r requirements.txt

3. Open and run the Jupyter Notebook:

   jupyter notebook CxValueSegmentation.ipynb

Dependencies

• Python 3.x
• Pandas
• NumPy
• Scikit-learn
• Matplotlib
• Seaborn

Visualizations

• RFM distribution across different clusters.
• Revenue distribution over time for different customer segments.
• Cluster analysis based on revenue metrics.

License

This project is licensed under the MIT License - see the LICENSE file for details.