What-if analysis is a great use case for Azure machine intelligence techniques. Bart Czernicki, Principal Technical Architect with the Microsoft Machine Intelligence team, shows you how in his sample web app for baseball decision analysis based on ML.NET and Blazor called the Machine Learning Workbench. He shows you an example architecture on Azure and provides all the source code on GitHub.
Bart has created the Machine Learning Workbench as a web app with a friendly interface to a powerful what-if analysis engine. It takes historical and current baseball data and uses AI and machine learning models to make informed predictions.
The solution delivers National Baseball Hall of Fame insights, but the architectural approach applies to decision analysis systems in general, from building a fantasy baseball team to forecasting financial scenarios for budgeting and planning.
The Machine Learning Workbench is built in ASP.NET core using ML.NET, an open-source framework that provides the inference engine, and Blazor Server to render the interface. Azure SignalR Service brokers communications between Workbench and the user interface.