Using AI for Analysis

Practical AI Techniques for Every Analysis Task

AI transforms data analysis from a technical specialty into an accessible skill. This chapter shows exactly how to use AI throughout your analysis workflow.

Uploading and Understanding Data

Getting Data to AI

Most AI assistants can work with data you:

  • Paste directly (smaller datasets)
  • Describe in detail
  • Upload as files (CSV, Excel)

First Look at Data

I've uploaded/pasted data about [topic].

Please:
1. Describe what columns/variables are present
2. Summarize the scope (how many records, time range, etc.)
3. Identify data types for each column
4. Note any obvious issues or missing data
5. Suggest what questions this data could answer

Data Profiling

Profile this dataset for me.

For each column, provide:
- Data type
- Count and missing values
- For numeric: min, max, mean, median
- For categorical: unique values and top categories
- Any notable patterns or issues

Asking Analytical Questions

Descriptive Questions

What are the total sales by [dimension]?
What's the average [metric] for [segment]?
How has [metric] changed over time?
What's the distribution of [variable]?

Comparative Questions

How do [segment A] and [segment B] compare on [metric]?
What changed between [period 1] and [period 2]?
Which [category] has the highest [metric]?
Are there significant differences between groups?

Diagnostic Questions

Why might [metric] have [changed/pattern]?
What factors are associated with [outcome]?
What explains the difference between [groups]?
What's driving [trend]?

Exploratory Questions

What patterns exist in this data?
What's surprising or unexpected?
Are there segments or clusters?
What relationships exist between variables?

Step-by-Step Analysis

Full Analysis Workflow

Help me analyze this data systematically.

Data: [Describe or provide]
Question: [What you want to understand]
Context: [Business context, how it will be used]

Walk me through:
1. Understanding the data
2. Cleaning and preparation needed
3. Exploratory analysis
4. Targeted analysis to answer the question
5. Interpretation and conclusions
6. Limitations and caveats

Specific Calculations

Calculate [specific metric] from this data.

Show me:
- The calculation itself
- The result
- How to interpret it
- Any concerns about validity

Visualization with AI

Requesting Charts

Create a [chart type] showing [what to display].

Data: [Provide data]
Key insight to highlight: [What should be obvious]
Audience: [Who will see this]

Visualization Recommendations

What's the best way to visualize this finding?

Finding: [What you discovered]
Data structure: [What you have]
Purpose: [Exploration, presentation, report]

Suggest options with pros/cons.

Interpreting Results

Understanding Findings

I found that [result].

Help me understand:
- What this means in practical terms
- Whether this is a large or meaningful effect
- What might explain this
- What I should investigate further
- How confident I should be

Sanity Checking

Does this result make sense?

My finding: [What you found]
Context: [What you'd expect]
Data: [What you analyzed]

Check for:
- Reasonableness
- Potential errors
- Alternative explanations

Writing Analysis Reports

Summarizing Findings

Help me summarize my analysis findings.

Key findings:
1. [Finding 1]
2. [Finding 2]
3. [Finding 3]

Audience: [Who will read this]
Format needed: [Email, report, presentation]
Level of detail: [Executive summary vs. detailed]

Create a clear summary that communicates insights effectively.

Creating Recommendations

Based on my analysis, what should we do?

Findings: [Summary of what you found]
Decision context: [What options exist]
Constraints: [Resources, timeline, etc.]

Help me develop actionable recommendations.

Troubleshooting Analysis

When Results Don't Make Sense

My analysis produced unexpected results.

What I found: [The unexpected result]
What I expected: [What should have happened]
My methodology: [How I did the analysis]

Help me figure out:
- Is the result actually wrong?
- What might explain it?
- How to verify?

Data Issues

I think there's something wrong with my data.

Symptom: [What seems off]
Data source: [Where it came from]
Expected: [What it should look like]

Help me diagnose the issue.

Building Analysis Skills

Learning Through Analysis

Walk me through this analysis step by step, explaining why we do each step.

Question: [What you want to learn]
Data: [What you have]

I want to understand the methodology, not just get results.

Practice Problems

Give me a data analysis practice problem.

Skill level: [Beginner/intermediate/advanced]
Topic: [Area to practice]
Type: [Exploratory, statistical, visualization]

Provide a scenario and dataset, then help me work through it.

What's Next

Analysis isn't complete until insights are communicated.

Next chapter: Communicating insights — turning analysis into action.