We trained each of the 1,600+ products in the client’s portfolio with a time series model to provide the monthly forecast for units sold over a 15-month rolling window. All this data was then written back to the database and presented in relevant Power BI dashboards.
The entire process of data ingestion, preparation and forecasting process was automated through a pipeline deployed using Azure Data Factory, Azure ML, and Azure Data Lake.
Power Apps and automated flows for users from Sales, Marketing, Manufacturing and Purchasing/Procurement teams to create fine tunings and validate forecast data.
With all this, our client now had an automated forecasting tool/framework which provides overall monthly sales forecasts for the next 15 months for over 1,600 products, and an automated approval flow that enables the various functions of the business to work together seamlessly.