Supply-Chain Demand Forecasting and Inventory Intelligence

An end-to-end planning system for a retail supply chain that runs on a student laptop CPU: three years of supply-chain data are loaded into a DuckDB SQL warehouse, a global LightGBM model forecasts the daily demand of 392 store-product series 28 days ahead with P10 / P50 / P90 quantiles, an inventory module turns the forecasts into safety stock, reorder points, order quantities and alerts, and a React BI application lets planners review and approve the orders. The data are synthetic (the fictional retailer Nila Mart).

Category: Machine Learning / Demand Forecasting / Inventory Optimisation / SQL Data Warehousing / Business Intelligence

CSE Project 36 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.