Summary
Auto-detection keys off column names. When a column is named unusually (e.g. gndr, patient_region_code), it can be missed or mis-typed. Let users manually map a column to a dimension (sex/race/age/geography) in the UI.
Why
Real datasets have messy headers; one wrong guess shouldnt silently drop a demographic axis.\n\n## Tasks\n- [ ] Show detected mapping with editable dropdowns per column\n- [ ] Re-run the engine on change\n- [ ] CLI equivalent: --map gndr=sex (track separately if scope grows)
Summary
Auto-detection keys off column names. When a column is named unusually (e.g.
gndr,patient_region_code), it can be missed or mis-typed. Let users manually map a column to a dimension (sex/race/age/geography) in the UI.Why
Real datasets have messy headers; one wrong guess shouldnt silently drop a demographic axis.\n\n## Tasks\n- [ ] Show detected mapping with editable dropdowns per column\n- [ ] Re-run the engine on change\n- [ ] CLI equivalent:
--map gndr=sex(track separately if scope grows)