Back to guides
2
4 min

Data Analysis with AI

Ask Questions About Your Data

From Spreadsheets to Insights

Most business professionals spend hours staring at spreadsheets trying to spot trends. AI can do this in seconds — if you know how to ask.

The key insight: AI does not "look" at your spreadsheet the way you do. When you paste CSV data into a prompt, the AI reads it as text — rows and columns become lines and commas. It then reasons about the patterns in that text. This means the quality of your analysis depends entirely on how you frame your questions.

Uploading Data

The simplest approach: copy your CSV data directly into the prompt. For small datasets (under 200 rows), this works well. For larger datasets, summarize first — paste the headers and a sample of rows, then describe the full dataset.

Good prompt:

> Here is our Q3 sales data (15 transactions). Analyze it and tell me:

> 1. Which region had the highest total revenue?

> 2. Which product sold the most units?

> 3. Are there any unusual patterns?

>

> [paste CSV data here]

Bad prompt:

> Look at this data and tell me something interesting.

The difference: specific questions get specific, useful answers. Vague prompts get vague, generic responses.

Finding Trends

AI is surprisingly good at spotting patterns humans miss — especially when comparing multiple dimensions at once:

  • Period-over-period changes — "Compare Q2 vs Q3 revenue by region"
  • Correlations — "Do higher-priced products sell fewer units?"
  • Outliers — "Flag any transactions where revenue is more than 2x the average"
  • Segment analysis — "Break down performance by product category and region"

    • Working with the Data

    When analyzing the pre-seeded sales data in this course, try these prompt patterns:

    Summary first: "Summarize this dataset — how many rows, what columns, what date range, any missing values?"

    Top/bottom: "What are the top 3 products by revenue? Bottom 3 by units sold?"

    Comparison: "Compare East vs West region performance. Which is growing faster?"

    Anomaly detection: "Are there any transactions that look unusual? Explain why."

    Recommendation: "Based on this data, which product-region combination should we invest in next quarter? Why?"

    Quality Checks

    AI analysis is only as good as the data. Before trusting any insight:

  • Verify the math — Ask AI to show its calculations. Check a few by hand.
  • Check the denominator — "Average revenue" means nothing without knowing how many transactions.
  • Question the sample — 15 rows is a tiny dataset. AI will find patterns, but they may not be statistically significant.
  • Look for what is missing — AI can only analyze what you give it. If your data excludes returns or discounts, the analysis is incomplete.

    • What You Will Build

    You will analyze a quarterly sales dataset using AI prompts. You will practice asking targeted questions, interpreting results, and verifying AI-generated insights against the source data.

    Glossary

    TermMeaning
    CSVComma-separated values — a simple spreadsheet format
    OutlierA data point significantly different from the rest
    Period-over-periodComparing the same metric across two time periods
    Segment analysisBreaking data into groups to compare performance
    Statistical significanceWhether a pattern is real or just random noise

    This is chapter 2 of AI for Business Decisions.

    Get the full hands-on course — free during early access. Build the complete system. Your projects become your portfolio.

    View course details
    Previous

    Ch. 1: What AI Can Do

    Next

    Ch. 3: Report Summarization