Summarize locally saved jobs and operational signals into trend-ready metrics without inventing production history.
The analytics module never seeds example history into the result. If this browser has no saved jobs, QA checkpoints or yield signal, the correct output is an empty state. That protects the operator from treating demonstration numbers as real production performance.
When history exists, read cost, yield and QA together but avoid causal shortcuts. A rise in spoilage and a rise in unit cost may share a cause, but the dashboard only shows that both changed. Investigation still requires the job records, material batch, equipment state and operator notes.
For meaningful trends, keep definitions stable. If ‘waste’ changes from press setup sheets to all non-shipped material halfway through the period, the percentages are not directly comparable until the definition is reconciled.
Metrics are calculated only from records actually saved in this browser. Empty history stays empty.
Waste, cost and QA can move together without one proving the other caused it. The dashboard keeps those signals separate.
A rising waste rate should trigger a process review, not an automatic diagnosis.