Retail Demand Forecasting
MGT 594 · Machine Learning · Tirthal Kothari
Retail demand modeling that connects data preparation, feature engineering, and replenishment planning.
The challenge
Our academic team worked with retail inventory, orders, pricing, and promotion data to forecast units sold and identify replenishment risk. The dataset needed cleaning and feature preparation before modeling.
What I did
I handled EDA and feature engineering: cleaning the dataset, encoding categorical variables, and creating features from price, discount depth, and inventory position. Our team then trained and compared regression models to forecast units sold.
The deliverable
A working forecasting pipeline and model comparison that surface potential SKU shortages to support replenishment planning.
Tools & methods
- Python
- pandas
- scikit-learn
- seaborn
- matplotlib
Project context
Academic team project in MGT 594. Deliverable: cleaned retail data, engineered features, and a forecasting model comparison for replenishment planning.