Demand Forecasting at Scale
A global CPG company had previously engaged Fractal to build an AI-enabled demand forecasting engine for one country. Now, the client wanted to scale the platform to multiple countries.
Fractal scientists brought the relevant files into Azure, storing the data in Blob. The files were loaded into SQL DB for processing (EDA/DI/DQ/feature engineering), where the data was used to run forecasting models, and the output stored back on the SQL DB. The model output was extracted as CSV files and transmitted back to the client.
Fractal established connectors to move the datasets into Azure and built a prototype to run the R-based forecasting engine on Azure ML studio. That prototype was then scaled to multiple country and category combinations.
Fractal developed a scalable forecasting engine on Azure for rapid deployment across multiple country and category combinations, delivering impactful, demand-based planning insights.