v1.2 depth expansion

Problem

After v1.0 stabilization, the profiler covers a baseline set of semantic types and statistical measures, but usage data will reveal which profile features analysts actually rely on and where they drop off. Some columns that users frequently inspect may lack useful profiling depth (e.g., JSON columns with no structure extraction, array columns with no element-level statistics), while other profile sections that were expensive to build may see little engagement. Without prioritizing depth expansion based on observed usage patterns and retention signals, engineering effort will be allocated by intuition rather than evidence, and the profiler will add features users don't need while leaving gaps in features they do.

Context

Possible Solutions

Plan

Implementation Progress

Review Feedback

• Review cleared