Prompt of the Day: Data Translator — Turn Tables and Numbers into Clear Insights
Monday morning, your boss sends you a spreadsheet with the quarterly numbers. 'Can you take a look and tell me what stands out?' 500 rows, 12 columns, no explanation. You stare at the numbers and don't know where to start.
Everyone knows this: reading data is easy, interpreting data is hard. What's a normal outlier, what's a real trend? Which number matters, which is just noise? This is exactly where AI becomes your data translator.
This prompt transforms raw data into a structured analysis — with the key insights, anomalies, and concrete recommendations you can drop straight into your presentation or email.
How to use the prompt:
1. Copy your data directly into the chat (table, CSV, or even a description of the numbers)
2. Describe the context — what is it about? What's being measured? What was the goal?
3. State who the analysis is for — your boss needs different insights than your team
4. Review the results critically: AI can confuse correlation with causation and invent connections
Why this works: AI spots patterns in numbers faster than you can — especially in large datasets. But the real advantage is the structure: instead of scrolling through the spreadsheet aimlessly, you immediately get the key points, sorted by relevance.
Pro tip: Upload your data as a CSV file (in ChatGPT, Claude, or Gemini) instead of pasting it as text. This works much better for larger datasets. For sensitive business data: check your company's privacy policies first — not all AI providers handle data the same way.
Important: AI analyses are a starting point, not a final product. AI can misinterpret numbers, miss outliers, or invent trends that don't exist. Use the analysis as a foundation and manually verify the key findings — especially when business decisions depend on them.
You are an experienced data analyst. I am giving you a dataset and need a clear, understandable analysis — no jargon, but insights I can share directly. **Context:** - What is being measured: [e.g. monthly sales figures, website traffic, customer satisfaction, project costs] - Time period: [e.g. January to March 2026, last quarter, last 12 months] - Goal/benchmark: [e.g. 10% growth year-over-year, budget of 50,000 EUR, NPS above 40] - Audience for the analysis: [e.g. executive leadership, my team, client, myself] **My data:** ''' [Paste table, CSV data, or description of numbers here] ''' **Please create the following analysis:** 1. **Executive Summary** (3-5 sentences): What are the key findings? What's going well, what isn't? Summarize as if you had 30 seconds to explain it to your boss in an elevator. 2. **Top 5 Insights:** The five most important data points or trends — each with: - What: What do the data specifically show? (with numbers) - Why it matters: What does this mean for us? - Recommended action: What should we do? 3. **Anomalies and Outliers:** Are there values that fall outside the norm? Unusual spikes or drops? List them and suggest possible explanations — but clearly mark assumptions with [hypothesis]. 4. **Trend Analysis:** What trajectory do the data show over the time period? Is there a clear upward/downward trend? Seasonal patterns? Turning points? 5. **Comparison with Goal/Benchmark:** Where do we stand relative to the goal? On track or not? What would need to change to reach the target? 6. **3 Concrete Recommendations:** What would you do next based on this data? Prioritized by impact. Each recommendation with a concrete first step. 7. **Visualization Suggestions:** Which 2-3 charts would best convey the key findings? (e.g. line chart for trends, bar chart for comparisons, heatmap for distributions) Briefly describe what goes on which axis. **Rules:** - Write so that someone WITHOUT statistics knowledge can understand the analysis - Use concrete numbers and percentages — no vague statements like 'significantly increased' - Clearly distinguish between **facts** (what the data show) and **interpretations** (what it might mean) - Mark uncertain conclusions with [hypothesis] - If the data are insufficient for a reliable conclusion, say so explicitly - Do NOT invent data points or numbers that are not in my data