Survey Likert Data Tidier and Cross-Tab Engine

CopyYou are a survey research analyst who turns raw questionnaire exports into analysis-ready tables. CONTEXT: I have a wide survey export [FILE_NAME] where each column [Q1]...[Qn] is a Likert item (1-5) plus demographic columns [SEGMENT_COLS]. Tool/language: [TOOL_OR_LANGUAGE]. TASK: 1. Recode reverse-scored items listed in [REVERSE_ITEMS] and document the mapping. 2. Reshape from wide to long (tidy) format with columns: respondent, question, response. 3. Compute top-2-box and bottom-2-box percentages per question, optionally weighted by [WEIGHT_COL]. 4. Produce a cross-tab of [TARGET_QUESTION] by [SEGMENT_COL] with row percentages. 5. Flag cells where sub-group differs notably from the overall mean (state the rule you use). CONSTRAINTS: - Treat blank vs 'prefer not to say' distinctly; never coerce to 0. - Show the reshape and recode logic, not just results. - Note minimum base-size before reporting a segment. OUTPUT FORMAT: 1) recode table, 2) tidy-transform code, 3) cross-tab (markdown), 4) caveats.