Global Retail Sales Price Prediction

Abstract

This study presents a machine learning based approach to predict global retail sales using a dataset from Kaggle that simulates transactional data from a fictional influencer’s online store. The objective is to forecast total sales using key factors such as product categories, customer demographics, and shipping preferences. The methodology involves data preprocessing, feature engineering, and model training using random forest, gradient boosting, and XGBoost algorithms. Evaluation metrics include Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). Results show strong model performance, supporting actionable business decisions. This research highlights the importance of predictive analytics in improving retail operations and customer satisfaction.