Gemini Prompts for Data Analysis
Google Gemini can work directly with uploaded files, making it great for data analysis tasks. These prompts show you how to ask Gemini to analyze spreadsheets, identify trends, suggest visualizations, and draft executive summaries—all while keeping the AI grounded in your actual data.
Last updated · By the Prompt Orange team
Prompt examples you can copy & use
Analyse this spreadsheet
“What's in this file?”
Too vague—AI has to guess what you want
“I've uploaded a CSV of monthly sales data by region. Analyze it and provide: (1) summary statistics (mean, median, range), (2) top 3 trends or patterns, (3) any anomalies or outliers, (4) 2 hypotheses for why Q2 dipped. Present as: Overview | Key findings | Hypotheses.”
Specific, clear, ready to use
Explain trends
“Why is this happening?”
Too vague—AI has to guess what you want
“Based on the uploaded sales data, explain why revenue increased 40% in the North region but stayed flat elsewhere. Consider: seasonality, marketing spend, product mix. Suggest 2 follow-up analyses we should run to confirm the cause.”
Specific, clear, ready to use
Draft an exec summary
“Summarize this data”
Too vague—AI has to guess what you want
“Create a 1-page executive summary of the attached Q3 performance data for our CEO. Include: 3-sentence TL;DR, key metrics vs. targets (table), biggest win, biggest concern, and 1 recommended action. Tone: clear, data-driven, no fluff. Max 300 words.”
Specific, clear, ready to use
Suggest visualisations
“What chart should I make?”
Too vague—AI has to guess what you want
“Based on the uploaded dataset (monthly user signups by source over 12 months), suggest 3 different chart types that would tell a clear story. For each, explain: what it shows, why it's effective, and what axes/groupings to use.”
Specific, clear, ready to use
Tips for using Gemini for data analysis
Upload the file directly rather than pasting data—Gemini handles file uploads well
Ask Gemini to 'think step by step' for complex analyses to improve reasoning
Request structured output formats: tables, bullet points, or specific sections
Be explicit about your audience (e.g., 'explain for a non-technical stakeholder')
Ask for limitations or caveats in the analysis to avoid overconfident conclusions