Customer Segmentation Using RFM and K-Means Clustering

1. Project Overview

This project applies data-driven customer segmentation using RFM (Recency, Frequency, Monetary) analysis combined with K-Means clustering.
The objective is to identify and categorize customer behavioral patterns from transaction data of a retail Superstore dataset, enabling targeted marketing, retention strategies, and personalized engagement.

The approach operationalizes advanced analytics and visualization to support data-driven decision-making in customer relationship management (CRM).

2. Business Objective

Organizations frequently struggle to allocate marketing resources efficiently due to limited visibility into customer purchase behavior.
This project demonstrates how data science can optimize marketing investment by identifying segments such as:

• Best / VIP Customers – High-value and high-frequency spenders.
• Loyal Customers – Regular buyers with strong brand retention.
• Potential / Regular Customers – Moderate value, high opportunity for upselling.
• At-Risk / Lost Customers – Previously active customers showing churn indicators.

The resulting segmentation drives personalized marketing campaigns, customer lifetime value prediction, and churn mitigation.

3. Dataset Description

The dataset used is a sample of transactional data from a global retail superstore.
Each record represents a unique purchase event, containing the following key attributes:

Field Name	Description
CustomerID	Unique customer identifier
InvoiceNo	Transaction ID
InvoiceDate	Date of purchase
Quantity	Number of items purchased
UnitPrice	Price per item
TotalAmount	Derived feature (Quantity × UnitPrice)

Data cleaning and aggregation were performed to ensure customer-level summarization for RFM modeling.

4. Methodology

Step 1: Data Preprocessing

• Handled null values and removed non-relevant transactions (e.g., returns or negative values).
• Aggregated transactions by customer to compute total spend, transaction count, and last purchase date.
• Calculated RFM metrics:
  • Recency = Days since last purchase
  • Frequency = Count of transactions
  • Monetary = Total revenue contributed

Step 2: Feature Scaling and Outlier Handling

• Applied IQR filtering for outlier reduction.
• Standardized RFM variables using StandardScaler to ensure balanced feature weighting before clustering.

Step 3: K-Means Clustering

• Employed the Elbow Method and Silhouette Score to determine the optimal number of clusters.
• Cluster centroids were interpreted as segment profiles.

Step 4: Segmentation Labeling

• Clusters were analyzed and mapped into business-relevant segments:
  • High Monetary + Low Recency + High Frequency → Best / VIP Customers
  • Moderate Monetary + Medium Recency → Loyal Customers
  • Low Frequency + Low Monetary + High Recency → At-Risk / Lost Customers

5. Visualization and Dashboard

A Power BI dashboard consolidates analytical insights for executive-level interpretation.
Visual components include:

• Recency vs Monetary Scatter Plot: Cluster distribution analysis
• Bar Charts: Frequency and Monetary contribution by segment
• Donut Chart: Revenue share per segment
• Summary Table: Average RFM values and recommended marketing actions

The dashboard integrates color-coded segmentation to support intuitive data storytelling and decision-making.

6. Model Evaluation

• Cluster stability and separation were validated through Silhouette Coefficient and intra-cluster variance analysis.
• The resulting clusters showed strong differentiation in behavioral metrics, confirming the suitability of K-Means for this RFM dataset.

7. Insights and Strategic Recommendations

Segment	Insight	Strategic Action
Best / VIP Customers	High spenders with frequent purchases	Reward and retain through exclusives and loyalty benefits
Loyal Customers	Consistent engagement but moderate spend	Encourage referrals and upsell programs
Potential / Regular Customers	Stable group with growth potential	Target with offers and personalized campaigns
At-Risk / Lost Customers	Low frequency and recent inactivity	Re-engage through win-back and feedback campaigns

This segmentation framework enables precision marketing, resource optimization, and customer lifetime value enhancement.

8. Technical Stack

• Programming: Python (Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn)
• Modeling: RFM Analysis, K-Means Clustering
• Visualization: Power BI (interactive dashboard)
• Versioning: Jupyter Notebook (.ipynb)
• Environment: Python 3.10+

9. How to Run

1. Clone the repository:

   git clone https://github.com//Customer-Segmentation-by-RFM-Kmeans.git cd Customer-Segmentation-by-RFM-Kmeans Install dependencies:

pip install -r requirements.txt Run the notebook:

jupyter notebook RFM_KMeans.ipynb Load the processed dataset into Power BI to reproduce the dashboard visuals.

10. Future Enhancements Integrate predictive modeling to forecast segment migration over time.

Develop automated Power BI refresh pipeline using Python connectors.

Extend segmentation with demographic and behavioral attributes for multidimensional analysis.

Integrate recommendation engine for personalized product suggestions.

Next Step for You:

1. In VS Code or Jupyter file explorer, create a new file → name it README.md.
2. Paste the above content inside it.
3. Save → commit → push to GitHub.