Objective Assume that you have the following datasets: • sales.csv, which is basically our train set. “customer_id”: customer identifier “month”: period when sales happened “material_id”: sold product identifier “amount”: product quantity sold
• customer.csv, which contains more details about our customer “customer_id”: customer identifier “size”: size of the customer (big, small, large, etc.,) “type”: type of the customer’s business (convenience store, bar, etc.,)
• products.csv, which contains more details about the portfolio of our products “product_id”: sold product identifier “category”: category of our products (ProductA, ProductB, etc.,)
• test.csv, which is the dataset that we are trying to predict. “customer_id”: customer identifier “month”: period when sales happened “category”: category of our products (ProductA, ProductB, etc.,) “amount”: actual amount sold “baseline prediction”: results obtained from a model developed by another data scientist
The purpose here is to help us develop a prediction model that enables us to predict the “test.csv” values as accurate as possible.