Time Series Forecasting • Demand Analytics • Python • Business Analytics

🎯 Objective

To build a reliable short-term forecasting framework capable of predicting weekly product demand.

Specifically to:

• Evaluate multiple forecasting approaches using historical demand data
• Compare model performance using walk-forward validation
• Select the most suitable model for each product category
• Generate forward forecasts for operational planning
• Translate forecast outputs into actionable business insights

🧠 Background

Demand forecasting plays a critical role in supply chain planning, inventory management, and production scheduling. Businesses must anticipate future demand patterns to minimise stockouts, reduce excess inventory, and optimise operational efficiency.

This project explores demand forecasting for a mid-sized FMCG company with three product categories: bakery products, beverages, and ready meals. Using historical weekly sales data, multiple forecasting approaches were evaluated to identify the most suitable model for each product category and generate short-term demand forecasts.

The project demonstrates how structured time-series modelling can be used to support data-driven operational decisions.

🧰 Tools & Methods

Tools:

1. Python
   1. Pandas
   2. Statsmodels
2. Excel