Al based demand forecast application

The paretos AI based Demand Forecasting App, built on Snowflake, offers high-precision forecasting capabilities with an easy and user-friendly data auto-connect. The app enables users to test how good forecasts can potentially be given the provided data. As clear data schemas are provided it furthermore enables users to experiment and test hypothesis on how other data inputs might increase forecast accuracy. The app is easy to use, delivers an exceptional forecast accuracy to enhance decision-making and operational efficiency across businesses.

Expected Workflow:
Two tables have to be filled with data first. One which include the truth timeseries to be predicted and another one providing meta data. Both exact schemas are defined in the documentation. Once the tables are filled they can be validated in the app if they are complete. As a next step the user sets up a direct share with the app provider for the two mentioned tables (Guided workflow is included in the app). Once the connection has been setup the user can start the training pipeline process. After the model training is finished and forcasts have been generated the user gets an automated notification via. email and can consume the first forecasts within a new shared database from the provider. From this point on the forecasts are updated in the database based on the schedule the user defined.