📊 CSV Analysis Prompt ⭐
"Analyze this CSV [paste or upload]: - Total rows, columns - Data types per column - Missing values per column - Outliers (top 10 by z-score) - Correlation matrix - 3 interesting patterns - 5 questions worth exploring Use ChatGPT Code Interpreter / Claude Analysis tool."
📈 Excel/Sheets Formula
"Convert this requirement to formula: 'For each row in column A, find the most recent date in column B where C matches A AND D > 100. Return "None" if no match.' Provide: - Excel formula - Google Sheets formula (if different) - Plain English explanation - Example with sample data"
🗃️ SQL Query Generation
"Schema: customers(id, name, country, signup_date, plan) orders(id, customer_id, amount, status, created_at) Question: 'Top 10 customers by total spend last 90 days, Pro plan, APAC region, minimum 5 orders.' Provide: - SQL (PostgreSQL) - Index recommendation - Complexity estimate - Alternative if data huge"
🧹 Data Cleaning
"Clean this messy data: [paste] Fix: - Inconsistent date formats - Phone numbers (spaces/hyphens) - Email casing - Trim whitespace - Standardize country (US, USA → United States) Provide: cleaned CSV + pandas script"
🐼 Python Pandas Prompt
"df columns: date, customer_id, amount Calculate: - Monthly revenue per customer - Cohort retention (by signup month) - Top 5 churners (biggest spend drop) Pandas code with comments + sample output."
📉 Visualization Prompt
"Recommend chart for: [data, question] - Why this chart type - Matplotlib/Plotly code - Color palette - Title, axis label - Annotation for insights"
🧮 Statistical Analysis
"A/B test result: Control: 5000 visit, 250 conversion Variant: 5000 visit, 290 conversion Calculate: - Conversion rate - Lift % - Statistical significance (chi-square) - Confidence interval - Sample size adequacy - Plain English verdict"
📊 Executive Report
"Create executive summary: [paste analysis] Format: - Headline finding (1 sentence) - 3 key metrics (number + change %) - 3 actionable recommendation - 1-2 risk/caveat - Length: half page"
🛠️ Code Interpreter Tool
- ChatGPT Advanced Data Analysis
- Claude Analysis tool
- Gemini Data Analyst
- CSV/Excel directly upload
- Sandboxed Python execute
- Chart, file generate
- Iterative — follow-up ask
🔄 Workflow Example
1. Upload sales.csv 2. "Profile: types, missing, distribution" 3. "Plot revenue by month, color by category" 4. "Anomaly identify — Q3 dip investigate" 5. "Forecast next 3 months with confidence interval" 6. "Generate slide deck summary"
⚠️ Privacy Warning ⭐
- Consumer ChatGPT-তে sensitive customer data upload করবেন না
- Enterprise/Team plan (no training) use করুন
- PII anonymize করুন আগে
- Local analysis (Jupyter + local LLM) if highly sensitive
💡 Pro Tips
- Column name descriptive হলে AI-এর analysis ভালো হয়
- Sample 20 rows paste → AI structure বুঝে
- "Show your work" — AI-এর reasoning দেখুন
- Chart re-generate "make it professional" বলে
- Complex query: AI-কে plan করতে বলুন আগে, execute পরে