🧠 Customer Segmentation using RFM and K-Means

Customer segmentation is a powerful way for businesses to understand and group their customers based on behavior. This project performs RFM Analysis (Recency, Frequency, Monetary) and uses K-Means Clustering to discover unique customer segments from retail transaction data.

---

📦 Dataset

• File: Online Retail.xlsx
• Description: Real-world online retail dataset containing transactions from 2010–2011.
• Key Fields: CustomerID, InvoiceNo, InvoiceDate, Quantity, UnitPrice

---

🔁 Workflow Overview

🔹 1. Data Preprocessing

• Removed rows with missing CustomerID
• Filtered out transactions with negative Quantity or UnitPrice
• Created a new column: TotalPrice = Quantity × UnitPrice

🔹 2. RFM Feature Engineering

• Recency: Days since last purchase (from the most recent date)
• Frequency: Number of unique purchases (InvoiceNo)
• Monetary: Total spending per customer

🔹 3. Exploratory Data Analysis

🔍 Pairplot of RFM Features

📊 Correlation Heatmap

---

🔹 4. Clustering

📈 Elbow Method to Select Optimal K

🟣 Silhouette Score Visualization

• Optimal number of clusters (k) is chosen based on lowest inertia and silhouette evaluation.

---

🔹 5. Final Pipeline

• Created a reusable scikit-learn pipeline with:
  • StandardScaler
  • KMeans
• Trained on the RFM features and predicted segments across the dataset.

---

📤 Output

• customer_segments.csv:
  • Contains: CustomerID, RFM metrics, assigned cluster
• Use these segments to identify:
  • VIP customers
  • At-risk/dormant users
  • Frequent but low spenders
  • New or seasonal shoppers

---

🛠️ Tech Stack

• Python 3
• pandas, numpy, matplotlib, seaborn, yellowbrick
• scikit-learn for clustering and pipeline
• jupyter for interactive development

---

🚀 Getting Started

1. Clone the repo

   git clone https://github.com/yourusername/customer-segmentation.git
   cd customer-segmentation
   

2. Install dependencies

    pip install -r requirements.txt
   

3. Place dataset in the project root.

4. Run the script to generate visualizations and the final csv.

   python main.py