Quality Analysis¶
ANALYSING THE DATA¶
Quality Analysis does for rejected units what Downtime Analysis does for stoppages. It groups bad production by the quality reason recorded against it, so losses can be broken down by cause rather than only counted.
Quality reasons are reasons of type Quality, managed on the Reasons admin page and scoped per station by the same include/exclude cascade as downtime reasons. A reason is attached to bad production in two ways:
- From the tablet. When an operator records a bad quantity, they are offered a quality reason to go with it. The pick is optional and can be skipped. Undoing an entry nets against the same reason. The free-form correction screen records no reason.
- From a device. A
bad_production_reasonfield in the same MQTT message as the bad count — see MQTT.
Bad production without a reason lands in an Unknown bucket. A reason code that is not (or no longer) configured on the account is still recorded and is shown as the raw code, so an attribution is never lost because configuration lagged behind.
The table drills down by the reason tree and shows, for each reason, the bad count and its cost (units × the SKU's value). Filter by station group, station, time period and a single SKU; export the selection to CSV.
The Production Entries page has a matching Bad Quality view listing each entry's reason, and the OEE Analysis page deliberately shows no reason breakdown — quality by reason lives here.
Recording bad quantities on the tablet is enabled per station with the Allow Reject Production flag; there is no separate flag for reasons.