Data Analysis Prompts

Prompts for exploring datasets, cleaning data, choosing the right analysis and explaining results. They make the AI state its assumptions and show its working before it draws conclusions.

• Data Analysis Report: Turn raw data into actionable insights.
• Customer Feedback Analysis: Analyze customer feedback and extract actionable insights.
• SQL Query Builder: Write a correct, readable SQL query and explain its logic.
• Dashboard Design Spec: Specify a dashboard around decisions rather than available metrics.
• A/B Test Results Analysis: Interpret experiment results honestly, including inconclusive ones.
• Cohort Analysis Plan: Analyse retention by cohort to see whether the product is improving.
• Data Cleaning Checklist: Clean a dataset methodically and document every decision.
• Forecast Model Brief: Build a forecast with stated assumptions and honest uncertainty.
• Metric Definition Document: Define a metric precisely enough that two teams compute it identically.
• Survey Analysis Plan: Analyse survey results without overstating what they support.
• Data Story Narrative: Turn an analysis into a narrative that drives a decision.
• Churn Analysis Framework: Find why customers leave and which causes are addressable.
• Spreadsheet Formula Helper: Build and explain a spreadsheet formula that survives editing.
• KPI Tree Builder: Decompose a top-line metric into the drivers teams can actually move.

Best data analysis prompts, in full

The top 3 by score on the free Verbito Prompt Scorer.

Customer Feedback Analysis (score 99/100)

You are a customer insights analyst specializing in turning feedback into actionable business strategies. Analyze the following customer feedback for [INSERT COMPANY/PRODUCT].

FEEDBACK TO ANALYZE:
[PASTE CUSTOMER REVIEWS, SUPPORT TICKETS, SURVEY RESPONSES]

ANALYSIS DELIVERABLES:

1. SENTIMENT OVERVIEW:
   - Overall sentiment breakdown (Positive/Neutral/Negative percentages)
   - Sentiment trend if data is time-series
   - Sentiment by category or source

2. THEME EXTRACTION:
   - Top 10 recurring themes
   - Frequency of each theme
   - Sentiment associated with each theme
   - Emerging themes (new in recent data)

3. PRAISE ANALYSIS:
   - What customers love most
   - Specific quotes representing each praise theme
   - Competitive advantages mentioned
   - Emotional language patterns

4. PAIN POINTS:
   - Top 10 complaints or issues
   - Severity rating (Critical/High/Medium/Low)
   - Frequency of each complaint
   - Impact on satisfaction/churn
   - Verbatim customer quotes

5. FEATURE REQUESTS:
   - Most requested features (ranked by frequency)
   - Feasibility assessment for each
   - Potential impact on satisfaction
   - Quick wins vs. strategic investments

6. COMPETITIVE MENTIONS:
   - Competitors mentioned
   - Comparison context (favorable/unfavorable)
   - Switching signals

7. ACTIONABLE RECOMMENDATIONS:
   - Immediate fixes (this sprint)
   - Short-term improvements (this quarter)
   - Strategic initiatives (this year)
   - Expected impact of each recommendation

8. VOICE OF CUSTOMER SUMMARY:
   - A narrative synthesis in the customer's voice
   - Key persona insights
   - Brand perception assessment

Audience: the product and support leads planning next quarter.
Output format: a table of themes with the count of mentions, one representative quote and a suggested action.
Constraints: only themes with at least 3 mentions; quote customers exactly.
Example row: "Slow exports | 14 | 'Exporting 2,000 rows takes minutes' | Add async export with email link".

Data Story Narrative (score 88/100)

You are an analyst who presents to executives and gets decisions made. Turn this analysis into a narrative: [INSERT FINDINGS].

Produce:

1. The headline finding in one sentence, leading with the implication rather than the method
2. Why this matters commercially, quantified
3. The three supporting points, each with one chart described precisely
4. The story arc: what we believed, what the data shows, what changes
5. The counter-argument and why it does not hold, addressed rather than omitted
6. The recommendation, specific and owned
7. What we still do not know and how much it matters
8. The one number the audience should remember
9. An appendix of methodology for those who will ask

Lead with the conclusion. Executives do not need the journey, and burying the finding on slide eleven loses the decision. Confidence should be proportionate to evidence: say likely when it is likely.

Tone: clear, confident, honest about uncertainty.

Data Cleaning Checklist (score 81/100)

You are a data scientist who documents cleaning decisions because they change conclusions. Build a cleaning plan for [INSERT DATASET] intended for [INSERT PURPOSE].

Work through:

1. Initial profile: row count, column types, memory, and a first look at each field
2. Missing values: extent, pattern, whether missingness is itself informative, and the handling decision with justification
3. Duplicates: how identity is defined, and how many exist
4. Outliers: detection method, and whether each is an error or a real extreme value
5. Type and format consistency: dates, currencies, categorical spellings
6. Referential integrity across joined sources
7. Units and scale consistency
8. Records excluded, with the count and reason for each exclusion rule
9. A before-and-after summary showing what the cleaning changed
10. The decisions a reviewer could reasonably disagree with, flagged

Never silently drop rows. Every exclusion is a decision that shapes the conclusion and must be visible in the output.

Tone: methodical and transparent.

How to use these prompts

1. Copy the prompt that matches your task.
2. Replace every [INSERT ...] part with your own details: the more specific, the better the answer.
3. Paste it into ChatGPT, Claude or Gemini. To tailor it to your exact situation, describe your goal in the Verbito prompt generator instead.

Build a data analysis prompt for your task · The 100 best prompts in one kit

Create a prompt or contact Verbito.