Frequency, severity, and recurrence signals across inspected parts, with special handling for PERFECT checks to separate clean outcomes from true drawing gaps.
Non-PERFECT findings
Systematic gap codes
PERFECT signals
Highest-burden part
Systematic signal
Part-level readout
Issue type ranking with cumulative contribution
Top-ranked codes are evaluated on non-PERFECT findings only so clean checks do not dilute the gap ranking.
Severity breadth by error code
Breadth is measured as how many parts each code touches; codes present in all three parts are treated as the strongest systematic signal.
Gap volume by recurrence class
Part-by-error heatmap
Filled cells indicate an observed signal for that part and code combination; PERFECT appears as a clean-result benchmark rather than a defect.