The FreshForecast solution is a big data platform to provide price dependent sales forecasts for fresh produce along the supply chain on regional and individual store levels. The platform is built on top of an existing, high-availability Spark cluster, designed to work with massive amounts of data, compatible with GS1 / EPCIS standards and will be connected to MIGROS data infrastructure. FreshForecast combines the internal data, provided by Migros, with external data from weather history and forecast, points of interest (competitors in the vicinity of stores) and seasonality.
FreshForecast provides:
- Trained prediction models powered by a deep learning algorithm
- A Big Data platform for continuous operation, curation and ML data pipelines and keeping the models accurate and returning suitable results to MIGROS as well as
- User Interfaces tailored to the needs of large-scale corporate operations, buying, but also for local store managers in order to enable decisions that sustainably reduce food waste in daily operation.