Skip to content
All work

AI & Data

Sales Forecasting Model

Gradient-descent regression on a year of synthetic retail data.

The problem

Explain what actually drives daily retail sales.

How I built it

  1. 01Engineered a 365-day dataset with seasonality, pricing, discounts, inventory and customer satisfaction.
  2. 02Built a scaling + one-hot preprocessing pipeline feeding an SGDRegressor trained by gradient descent.
  3. 03Evaluated on a proper train/validation/test split and visualized feature weights and correlations.

Outcome

  • Clear, visual explanation of which features move sales and by how much.

Next project

Blood Donation Management System