ChatGPT Data Analysis & Insight Generator Prompt
Turn raw data into clear insights, patterns, and actionable recommendations.
Category
💻 Coding
Difficulty
Intermediate
Models
3
Last Updated
2026-06-28
Works with
📄 Example output
⚠️ Common Mistakes
❓ FAQ
⚙️ Fill in your variables
📋 Prompt
You are a senior data analyst who has built reporting systems and extracted insights from complex datasets for Fortune 500 companies.
Data description: [data description — what this data is and what it measures]
Business context: [business context — what decision this data supports]
Analysis goal: [analysis goal — identify trends/find anomalies/compare periods/find correlations/build a report]
Paste your data:
[paste data here — CSV, table, or description]
Task: Complete data analysis:
1. DATA SUMMARY:
- What you have: rows, columns, date range, key metrics
- Data quality issues: missing values, outliers, suspicious entries
2. KEY FINDINGS (5–7 insights, ranked by business importance):
For each: Finding → Evidence (the specific numbers) → Business implication
3. ANOMALIES & FLAGS:
- Any data points that require investigation
- Possible explanations for each
4. TREND ANALYSIS:
- Direction and magnitude of change over time
- Seasonality or cyclical patterns if applicable
- Leading indicators (what predicts future performance)
5. ACTIONABLE RECOMMENDATIONS:
3–5 specific actions supported by the data
6. WHAT THE DATA CANNOT TELL YOU:
Limitations and what additional data would improve the analysis
7. VISUALISATION SUGGESTIONS:
The 3 charts that would best communicate these findings
Data description: [data description — what this data is and what it measures]
Business context: [business context — what decision this data supports]
Analysis goal: [analysis goal — identify trends/find anomalies/compare periods/find correlations/build a report]
Paste your data:
[paste data here — CSV, table, or description]
Task: Complete data analysis:
1. DATA SUMMARY:
- What you have: rows, columns, date range, key metrics
- Data quality issues: missing values, outliers, suspicious entries
2. KEY FINDINGS (5–7 insights, ranked by business importance):
For each: Finding → Evidence (the specific numbers) → Business implication
3. ANOMALIES & FLAGS:
- Any data points that require investigation
- Possible explanations for each
4. TREND ANALYSIS:
- Direction and magnitude of change over time
- Seasonality or cyclical patterns if applicable
- Leading indicators (what predicts future performance)
5. ACTIONABLE RECOMMENDATIONS:
3–5 specific actions supported by the data
6. WHAT THE DATA CANNOT TELL YOU:
Limitations and what additional data would improve the analysis
7. VISUALISATION SUGGESTIONS:
The 3 charts that would best communicate these findings
DATA SUMMARY:
Dataset: 6 months (Jan–Jun 2026), 8 metrics, 24 data points
Key metrics: Sessions, Page views, Bounce rate, Time on site, Conversions, Conversion rate, Source, Content type
Data quality: No missing values. June data is partial (21 days) — flag in analysis.
KEY FINDINGS:
1. CONVERSION RATE DISPARITY BY CONTENT TYPE (High business impact)
Prompt pages convert at 4.2% vs. calculator tools at 1.8% vs. blog posts at 0.4%.
Implication: Prompt pages are 2.3× better at converting than calculators and 10.5× better than blog posts. Priority: maximise prompt page count and organic entry points.
2. TRAFFIC SPIKE IN MARCH NOT MATCHED BY CONVERSIONS (Flag for investigation)
March traffic increased 340% (likely Product Hunt launch based on referrer data) but conversion rate dropped to 0.9% (vs 3.1% average).
Implication: Product Hunt traffic doesn't convert — quality of traffic matters more than volume. Don't optimise for PH traffic specifically.
3. ORGANIC SEARCH TRAFFIC GROWING 23% MoM (Positive trend)
SEO traffic is the fastest-growing channel and has the highest conversion rate (5.1%).
Recommendation: Double down on SEO-optimised prompt pages.
ACTIONABLE RECOMMENDATIONS:
1. Publish 20 more prompt pages targeting high-intent SEO keywords (evidence: prompt pages convert 4.2% from organic)
2. Add internal links from calculator tools to relevant prompt pages (evidence: conversion rate disparity)
3. Investigate March spike source — if Product Hunt, don't repeat for conversion goals
Dataset: 6 months (Jan–Jun 2026), 8 metrics, 24 data points
Key metrics: Sessions, Page views, Bounce rate, Time on site, Conversions, Conversion rate, Source, Content type
Data quality: No missing values. June data is partial (21 days) — flag in analysis.
KEY FINDINGS:
1. CONVERSION RATE DISPARITY BY CONTENT TYPE (High business impact)
Prompt pages convert at 4.2% vs. calculator tools at 1.8% vs. blog posts at 0.4%.
Implication: Prompt pages are 2.3× better at converting than calculators and 10.5× better than blog posts. Priority: maximise prompt page count and organic entry points.
2. TRAFFIC SPIKE IN MARCH NOT MATCHED BY CONVERSIONS (Flag for investigation)
March traffic increased 340% (likely Product Hunt launch based on referrer data) but conversion rate dropped to 0.9% (vs 3.1% average).
Implication: Product Hunt traffic doesn't convert — quality of traffic matters more than volume. Don't optimise for PH traffic specifically.
3. ORGANIC SEARCH TRAFFIC GROWING 23% MoM (Positive trend)
SEO traffic is the fastest-growing channel and has the highest conversion rate (5.1%).
Recommendation: Double down on SEO-optimised prompt pages.
ACTIONABLE RECOMMENDATIONS:
1. Publish 20 more prompt pages targeting high-intent SEO keywords (evidence: prompt pages convert 4.2% from organic)
2. Add internal links from calculator tools to relevant prompt pages (evidence: conversion rate disparity)
3. Investigate March spike source — if Product Hunt, don't repeat for conversion goals
🏆
💡 Pro Tips
Best model for this prompt
Claude
Claude (Opus 4 / Sonnet 4)
Always state what the data cannot tell you — analysis without acknowledged limitations is misleading
Rank findings by business importance, not statistical significance — a 0.1% change that affects $1M of revenue matters more than a 50% change that affects $100
Every finding should connect to a decision — 'traffic increased 23%' is a fact; 'traffic increased 23%, suggesting our SEO investment is paying off and we should double it' is an insight
Be specific about the numbers behind every finding — 'conversion rate improved' is weak; 'conversion rate improved from 1.8% to 3.1% over 3 months' is useful
Reporting data instead of analysing it — a table of numbers is not an analysis; pattern recognition and implication are
Treating correlation as causation — traffic increased and so did conversions is not evidence that traffic caused conversions
Over-indexing on outliers — one anomalous month doesn't change the trend; be careful about drawing conclusions from single data points
Ignoring data quality issues — analysis built on bad data produces confident wrong answers
- How much data do I need for meaningful analysis?For trend analysis: minimum 12 data points (12 months, or 12 weeks). For A/B testing: depends on traffic and effect size — use a sample size calculator. For finding patterns: more is better, but 6 months of weekly data is usually enough to identify seasonality and trends.
- What visualisations work best for different data types?Time series (performance over time): line chart. Comparison between categories: bar chart. Correlation between two variables: scatter plot. Part-to-whole: pie chart (max 5 segments) or stacked bar. Distribution: histogram. For most business reporting, line and bar charts cover 80% of needs.
- Which AI model is best for data analysis?For interpreting data you paste as text or a table: Claude and ChatGPT are both strong. For actual computation and statistical analysis: use Python (pandas, numpy) or R, then use AI to interpret the results. AI should interpret data, not replace statistical computation.
- Can I paste actual CSV data into the prompt?Yes — paste up to ~50 rows of data directly into the prompt. For larger datasets, summarise the data (key metrics, date range, totals by category) and paste that. For very large datasets, use Python or a BI tool to pre-process, then use AI to interpret the summary statistics.