Data Analysis Prompts for Business Intelligence Teams

AI can dramatically accelerate data analysis—from writing complex SQL queries to extracting insights from datasets. This guide provides proven prompts for data professionals at every level.

Why Data Analysis Prompts Matter

The right prompt turns vague requests like "analyze this data" into precise, actionable outputs. Knowing how to write better AI prompts makes all the difference. Good data prompts:

• Specify the data structure you're working with
• Define the question you're trying to answer
• Indicate the output format needed
• Set context for business relevance

Let's dive into the prompts.

SQL Query Prompts

Basic Query Generation

Write a SQL query to [GOAL].

Database: [PostgreSQL / MySQL / SQLite / etc.]

Schema:
[TABLE DEFINITIONS WITH COLUMNS AND TYPES]

Requirements:
- [Performance considerations]
- [Any constraints]

Include comments explaining complex parts.

Complex Multi-Table Query

I need a query that [BUSINESS QUESTION].

Tables involved:
- [table1]: [columns, relationships]
- [table2]: [columns, relationships]
- [table3]: [columns, relationships]

Expected output columns:
- [column1]: [description]
- [column2]: [description]

Filters:
- [Filter conditions]

Aggregations needed:
- [Any GROUP BY, SUM, AVG, etc.]

Optimize for [large dataset / real-time / one-time report].

Query Optimization

Optimize this SQL query for performance:

[QUERY]

Table statistics:
- [table1]: ~[N] rows
- [table2]: ~[N] rows

Current execution time: [TIME]
Target: [TARGET TIME]

Existing indexes:
[INDEX LIST]

Suggest:
1. Query rewrites
2. Index recommendations
3. Explain the optimization reasoning

Debugging SQL

This query returns wrong results:

[QUERY]

Expected: [DESCRIBE EXPECTED OUTPUT]
Actual: [DESCRIBE ACTUAL OUTPUT]

Schema:
[RELEVANT SCHEMA]

Sample data:
[SAMPLE DATA IF HELPFUL]

Find the bug and fix it.

Window Functions

I need to calculate [METRIC] using window functions.

Use case: [DESCRIBE WHAT YOU'RE TRYING TO DO]

For example: "Running total" or "Rank within group" or "Moving average"

Table:
[SCHEMA]

Show the query with explanation of each window function component.

Browse our data prompts collection for more SQL templates.

Data Exploration Prompts

Initial Data Assessment

I have a new dataset. Help me explore it:

Columns:
[LIST COLUMNS WITH TYPES]

Sample rows:
[PASTE A FEW ROWS]

For each column, suggest:
1. Data quality checks to run
2. Summary statistics to calculate
3. Visualizations to create
4. Potential relationships to explore

Finding Patterns

Analyze this data for patterns:

[DATA SAMPLE OR DESCRIPTION]

Look for:
- Trends over time
- Correlations between variables
- Anomalies or outliers
- Seasonal patterns
- Group differences

Present findings as [bullet points / narrative / structured report].

Cohort Analysis Setup

Help me set up a cohort analysis:

Business: [DESCRIPTION]
Cohort definition: [How to group users - signup date, first purchase, etc.]
Metric to track: [Retention, revenue, engagement, etc.]
Time periods: [Week / month / quarter]

Data available:
[TABLE AND COLUMN DESCRIPTIONS]

Provide:
1. SQL query to generate the cohort table
2. How to visualize the results
3. What to look for in the analysis

Funnel Analysis

Create a funnel analysis query:

Funnel steps:
1. [Step 1 - e.g., page view]
2. [Step 2 - e.g., add to cart]
3. [Step 3 - e.g., checkout started]
4. [Step 4 - e.g., purchase completed]

Events table schema:
[SCHEMA]

Calculate:
- Absolute numbers at each step
- Conversion rates between steps
- Overall funnel conversion
- Drop-off points

Insight Extraction Prompts

Business Insights from Data

I have this analysis result:

[DATA OR SUMMARY]

Business context: [WHAT THE BUSINESS DOES, GOALS]

Extract:
1. 3-5 key insights
2. For each insight:
   - What the data shows
   - Why it matters for the business
   - Recommended action
3. Questions this raises for further analysis

Comparing Periods

Compare these two periods:

Period 1 ([DATE RANGE]):
[METRICS]

Period 2 ([DATE RANGE]):
[METRICS]

For each metric:
1. Calculate the change (absolute and percentage)
2. Assess if change is significant
3. Suggest possible causes
4. Recommend next steps

Anomaly Explanation

I found an anomaly in our data:

What I observed: [DESCRIBE THE ANOMALY]
Normal range: [WHAT'S TYPICAL]
When it occurred: [DATE/TIME]

Business context: [RELEVANT BUSINESS INFO]
Recent changes: [ANY KNOWN CHANGES]

Help me:
1. List possible causes (ranked by likelihood)
2. Suggest data to examine to confirm each cause
3. Recommend immediate actions

Root Cause Analysis

Our [METRIC] dropped by [X%] on [DATE].

Available data:
- [Data source 1]
- [Data source 2]

Segmentations available:
- [Dimension 1]
- [Dimension 2]

Guide me through root cause analysis:
1. What segments to check first
2. Queries to run for each hypothesis
3. How to validate the root cause

Reporting Prompts

Executive Summary

Create an executive summary from this analysis:

[DETAILED ANALYSIS OR DATA]

Audience: [C-level / department heads / board]
Key business question: [WHAT THEY WANT TO KNOW]

Format:
- Headline (1 sentence takeaway)
- 3-4 bullet points with key findings
- Recommendation for action
- Risk or caveat if applicable

Length: Under 200 words
Avoid jargon—explain data implications in business terms.

Weekly Metrics Report

Generate our weekly metrics report narrative:

This week's numbers:
[METRICS WITH VALUES]

Last week:
[LAST WEEK'S VALUES]

MTD target:
[TARGETS]

Write a brief report that:
- Opens with the headline (best or worst news)
- Compares to last week
- Notes progress toward targets
- Highlights 1-2 items needing attention
- Ends with next week's focus

Tone: [Professional / casual / urgent]

Dashboard Description

Write descriptions for a dashboard I'm building:

Dashboard purpose: [WHO USES IT, FOR WHAT]

Metrics to describe:
1. [Metric name]: [Definition, why it matters]
2. [Metric name]: [Definition, why it matters]
3. [Metric name]: [Definition, why it matters]

For each metric, write:
- Short definition (1 sentence)
- How to interpret it (what's good/bad)
- How it's calculated

Stakeholder Presentation

Help me structure a presentation of this analysis:

Findings:
[YOUR KEY FINDINGS]

Audience: [WHO'S WATCHING]
Time available: [MINUTES]
Goal: [INFORM / PERSUADE / DECIDE]

Create:
1. Slide outline with key message per slide
2. Suggested visualizations
3. Talking points for each slide
4. Anticipated questions and answers

Visualization Prompts

Chart Selection

I want to visualize [DATA DESCRIPTION].

Goal: Show [RELATIONSHIP / TREND / COMPARISON / DISTRIBUTION]
Audience: [WHO WILL SEE THIS]

Suggest:
1. Best chart type and why
2. Alternative options
3. Axis labels and title
4. Color recommendations
5. Common mistakes to avoid

Chart Critique

Critique this visualization approach:

Data: [WHAT DATA]
Current chart: [CHART TYPE]
Goal: [WHAT IT'S TRYING TO SHOW]

Identify:
- What works well
- Potential misinterpretation risks
- Improvements to suggest
- Alternative approaches

Python Visualization Code

Generate Python code to create a [CHART TYPE]:

Data structure:
[DESCRIBE YOUR DATAFRAME]

Libraries: [matplotlib / seaborn / plotly]

Requirements:
- [Size specifications]
- [Color scheme]
- [Labels and titles]
- [Any annotations]

Include comments explaining customization options.

Data Cleaning Prompts

Data Quality Assessment

Review this data sample for quality issues:

[DATA SAMPLE]

Check for:
- Missing values patterns
- Inconsistent formats
- Duplicates
- Outliers
- Invalid values
- Referential integrity issues

Provide remediation suggestions for each issue found.

Cleaning Script Generation

Write a [Python / SQL] script to clean this data:

Current issues:
- [Issue 1]
- [Issue 2]
- [Issue 3]

Raw data format:
[SAMPLE]

Desired clean format:
[EXPECTED OUTPUT]

Include logging for tracking what was cleaned.

Standardization Rules

Create standardization rules for [FIELD TYPE]:

Examples of current messy data:
[SAMPLE DATA]

Business requirements:
[WHAT THE CLEAN DATA SHOULD LOOK LIKE]

Generate:
1. Mapping rules
2. Regex patterns if applicable
3. Edge case handling
4. Validation checks

A/B Testing Prompts

Experiment Design

Help me design an A/B test:

Goal: [WHAT WE WANT TO LEARN]
Primary metric: [KEY METRIC]
Baseline: [CURRENT VALUE]
Minimum detectable effect: [% CHANGE WE CARE ABOUT]
Traffic available: [USERS/DAY]

Determine:
1. Required sample size
2. Expected duration
3. Segmentation strategy
4. Guardrail metrics to track
5. Potential confounds to control

Results Analysis

Analyze these A/B test results:

Control: [N] users, [METRIC VALUE]
Treatment: [N] users, [METRIC VALUE]

Test duration: [TIME]
Expected effect: [WHAT WE HOPED FOR]

Provide:
1. Statistical significance (include method used)
2. Confidence interval
3. Practical significance assessment
4. Recommendation (ship / iterate / abandon)
5. Caveats or concerns

Frequently Asked Questions

How detailed should I be about my schema?

Include all tables and columns relevant to your question. More context = better queries. But you don't need the entire database—just what's needed.

Can AI write production-ready SQL?

Treat AI-generated SQL as a solid first draft. Always:

• Review the logic
• Test on sample data
• Check performance on large datasets
• Validate results against known values

What about sensitive data?

Never paste PII or confidential data into AI prompts. Use synthetic or anonymized samples. Describe the structure without real values.

Which AI is best for data work?

Claude handles complex multi-table logic well. ChatGPT with Code Interpreter can actually run Python/SQL. For production queries, always validate independently.

How do I get better at data prompts?

Study how senior analysts ask questions. The prompt structure mirrors how you'd brief a human analyst. Be specific about the business context—our business strategy prompts guide has templates tailored to strategic decisions.

---

Explore our complete data prompts library with 100+ templates for SQL, analytics, and visualization.