Analyze Workflow

Core loop

The Analyze page is designed around a repeatable loop:

1. choose lot and snapshot
2. find a suspicious test
3. inspect it in linked views
4. isolate a population by brushing or filtering
5. save or share the conclusion

Context that propagates

Most of the page shares a common context:

• selected lot
• selected snapshot
• current test
• wafer scope when applicable
• active die selection
• active scenario limits

That propagation is the reason to work from Analyze instead of exporting to external tools too early.

Cards vs. modals

• Cards are for quick pattern recognition and lightweight iteration.
• Modals are for the full interaction model and higher-resolution inspection.

If the card looks directionally useful but cramped, open the modal instead of forcing a deep read from the card.

Selection model

Brushing is shared state across the analysis surfaces.

• Shift + drag adds to the current selection
• Alt + Shift + drag replaces the current selection

Use selection when you want to answer “what is special about this subset?” rather than “what is the lot doing overall?”

Saving

Save when the notebook captures:

• a meaningful test focus
• a useful selection or wafer focus
• what-if limits worth reviewing with someone else

Do not save every exploratory click sequence. The notebook history is more useful when versions correspond to actual conclusions.