06Live
Retail Demand Forecasting
End-to-end machine learning demand forecasting system
Machine LearningForecastingDeployment
Stack
PythonPandasScikit-learnRandom ForestFlaskJupyter
Overview
A comprehensive data science project that predicts product demand from product, region, price, and calendar features. The project covers data validation, exploratory analysis, feature engineering, model comparison, deployment, automated testing, and business reporting.
Gallery
click to expand
Highlights
- Built an end-to-end demand forecasting pipeline with data validation, EDA, and feature engineering
- Compared Linear Regression, Random Forest, and Gradient Boosting, achieving 2.32-unit MAE with Random Forest
- Deployed the trained model as a Flask REST API with an interactive client-side dashboard
- Implemented automated pipeline and API testing with 8/8 deployment tests passing