Demand forecasting + replenishment

Favorita Planner Exception Queue

Which item/store/family combinations need replenishment review first, and what reasons justify each exception?

Decision summary

What this proves

A controlled Favorita slice becomes baseline forecasts, metric tables, ABC/XYZ segments, safety-stock assumptions, reorder-point assumptions, and a ranked planner exception queue.

  • WAPE, MAE, Bias %, Forecast Score, and FVA per item/store/family.
  • Planner queue ranked by demand volume, absolute bias, error, and variability.
  • Reason codes include high_volume, under_forecast_bias, over_forecast_bias, high_error, volatile_demand, and replenishment_risk.
  • Safety stock and reorder point are transparent planning assumptions, not purchase instructions.
Reviewer mode

Fast evidence review

Claims stay bounded by available public data. Outputs are decision-support candidates, not automated operational instructions or unsupported causal proof.