Back to prompts
Data AnalysisChatGPTClaudeGemini

Dataset Summary and Insights

Paste or describe a dataset and get an instant summary of key statistics, patterns, anomalies, and actionable insights.

Prompt Template

You are a senior data analyst. Analyze the following dataset and provide a comprehensive summary:

Dataset description: [what the data is about]
Source: [where it came from]
Time period: [date range]
Key columns: [list column names and types]
Business question: [what we're trying to answer]

[PASTE SAMPLE DATA OR DESCRIBE IT HERE]

Provide:
1. Dataset overview (rows, columns, data quality notes)
2. Key statistics for numeric columns (mean, median, range, outliers)
3. Top 5 findings / patterns
4. Anomalies or data quality issues
5. 3 actionable recommendations based on the data
6. Suggested next analyses to run

Example Output

Dataset Analysis: E-commerce Sales Q1 2026

**Overview:** 12,450 rows × 8 columns. Data appears clean with 0.3% null values in 'discount_applied' column (acceptable).

Key Statistics

| Metric | Value |

|--------|-------|

| Total Revenue | $2.4M |

| Avg Order Value | $192 |

| Median Order Value | $134 (right-skewed — a few large orders pull mean up) |

| Highest Revenue Category | Electronics (38%) |

| Return Rate | 11.2% |

Top 5 Findings

1. **Weekend orders have 23% higher AOV** than weekday orders

2. **February dip:** Revenue dropped 31% in week 2 — correlates with shipping delay incident

3. **Mobile conversion** (2.1%) is half of desktop (4.3%) — UX opportunity

4. **Repeat customers** (28% of base) generate 61% of revenue

5. **Returns highest in Clothing** (18.4%) — sizing issue likely

Recommendations

1. Investigate and fix mobile checkout friction (potential +$240K/quarter)

2. Launch retention campaign targeting the 72% of customers who've only ordered once

3. Review clothing size guides and add size charts to reduce returns

Tips for Best Results

  • 💡Paste a 10-20 row sample with column headers — even a small sample enables much better analysis than a description alone
  • 💡Always state the business question upfront — the same data yields different insights depending on what you're trying to decide
  • 💡Ask for a 'data quality scorecard' separately if your dataset is large — issues like nulls and duplicates need their own audit