Demand Forecasting using Machine Learning
A large, global distributor wanted to improve their demand forecasts and evaluate whether or not machine learning would improve their accuracy. The company purchases items from various manufacturers and then stocks the shelves of big-box retailers as well as smaller, independent shops with their products. Currently, the distributor utilizes a third-party forecasting solution to help decision makers make more accurate purchase orders from manufacturers.
Using a Cortana Analytics Machine Learning model in Azure, the distributor was able to increase the accuracy of their forecasts, reduce the excess product they keep on hand and reduce the amount of working capital invested in inventory.
Technologies: Cortana Analytics Machine Learning, Azure SQL DB, Power Pivot, Power View, Excel Industry: Consumer Product Goods, Manufacturing
Our Work Included:
- Extracting data from an on-prem SQL Server data warehouse
- Loading the data into an Azure SQL database
- Running the data through multiple Cortana Analytics machine learning models to determine the best model for the forecast
- Validate and visualize the results in Power Pivot, Power View and Excel