Machine Learning Framework for Customer Profiling and Segmentation Using Demographic and Behavioral Data

In the current era of data-driven decision-making, comprehending customer behavior is fundamental for businesses striving to deliver personalized services, strengthen engagement, and support strategic planning. This study presents the design and implementation of a machine learning–based framework for customer profiling using demographic and behavioral variables. The primary aim is to categorize customers into meaningful segments according to factors such as age, gender, profession, spending habits, family size, and marital status. These customer segments can help organizations tailor marketing strategies, thereby improving both operational efficiency and customer satisfaction. The system applies several supervised learning algorithms—Logistic Regression, K-Nearest Neighbors (KNN), Naive Bayes, Random Forest, Gradient Boosting, and particularly the Classification and Regression Tree (CART) model. CART was ultimately chosen due to its balance of simplicity, interpretability, and predictive performance. Model evaluation was conducted using accuracy, precision, recall, and F1-score metrics, while hyperparameter optimization was performed through GridSearchCV to enhance performance. To ensure accessibility and real-time use, the finalized model was deployed within an interactive Streamlit web application. Users can input customer details and immediately obtain the predicted customer category. The application also includes visualization components such as confusion matrices, feature importance plots, and segment distribution charts to facilitate interpretation. The architecture emphasizes modularity, scalability, and usability, supporting potential future integration with enterprise systems. This work highlights the effectiveness of machine learning in customer analytics and provides a practical, interpretable, and deployable tool for business decision-making.