Logo Lanfrica

abbhas2011/retail-customer-segmentation

Domain:

socioeconomic

Record type:

project
Creator:
abb
Host:
3MTT Capstone Project - Retail Customer Segmentation in Nigeria # retail-customer-segmentation # πŸ›’ Retail Customer Segmentation for Nigerian Businesses (3MTT Capstone) ## πŸ“Œ Project Overview In Nigeria, many small-to-medium retail and POS businesses treat all customers equally, using one-size-fits-all marketing strategies. This project uses machine learning to solve this over-generalization problem by segmenting customers based on their purchasing behaviors using transaction data. By implementing an **RFM (Recency, Frequency, Monetary)** model and **K-Means Clustering**, this solution allows local retailers to identify high-value buyers, retain at-risk customers, and drive targeted promotions. --- ## 🎯 MVP Features * **Data Processing & Cleaning:** Handles missing values, handles transaction logs, and filters invalid records. * **RFM Feature Engineering:** Converts transaction history into actionable metrics: * **Recency ($R$):** Days since last transaction. * **Frequency ($F$):** Total purchase count. * **Monetary ($M$):** Total spent in Naira ($\text{NGN}$). * **K-Means Clustering:** Groups customers into distinct, non-overlapping segments. * **Model Evaluation:** Silhouette Score and Calinski-Harabasz Index used to validate optimal cluster counts. * **Customer Persona Profiling:** Maps raw clusters to real-world business actions. --- ## πŸ› οΈ Tech Stack & Tools * **Language:** Python 3.x * **Libraries:** `pandas`, `numpy`, `scikit-learn`, `matplotlib`, `seaborn` * **Environment:** Jupyter Notebook / Google Colab --- ## πŸ“Š Key Results & Evaluation Metrics * **Optimal Clusters ($K$):** 3 * **Silhouette Score:** ~0.45+ (confirms well-separated clusters) * **Identified Customer Personas:** 1. **High-Value Champions:** Frequent buyers with high total spend. *Action: Enroll in VIP loyalty/cashback programs.* 2. **Regular Shoppers:** Consistent transaction frequency with moderate spend. *Action: Cross-sell related items.* 3. **At-Risk / Inactive Customers:** Haven't purchased recently. *Action: Send re-engagement discount SMS/Wha …