Versia – Activity prediction
Versia
Sector: Commercial sector
Business Case
This is an activity prediction algorithm. While the direct experience is in the field of retail, its applicability is universal. It is a question of demand forecasting.
Objectives
Match operational resources to expected demand (operations management).
Use case
A model was built that allows prediction using time series and the exceptions that activity calendars infer to demand. In addition, external data sources on weather forecasting were incorporated for application in very specific business contexts.
Infrastructure
Historical time series of activity by production centre. Private datasets.
Technology
Automatic or Deep Learning
Data
Historical time series of activity by production centre. Private datasets.
Resources
Tribu asignada al proyecto. Equipo multidisciplinar: Tic, Matemáticos/Físicos, Ing. Organización._x000B_GCP, AWS, HEROKU, VAULT.
Difficulties and learning
The training of demand exceptions generated by marketing campaigns. Real-time system interoperability.
KPIs (business impact and metrics of the model)
Margin for error on the real activity recorded.
Funding
Private. Successful.
Collaborators, Partners
Own resources.
