Analysing Data with AI: From Spreadsheets to Insights

AI is increasingly capable of working directly with data — not just text. This lesson covers the practical techniques for using AI to analyse data, from simple spreadsheet work to more complex analytical tasks.

The Two Main Approaches

Approach 1: Describe your data, ask for analysis frameworks Tell the AI what data you have and ask it how to analyse it.

"I have a spreadsheet with monthly sales data broken down by: region, product category, sales rep, and channel. I want to understand what is driving the 15% decline in the North region over the last quarter. What analyses should I run? What should I look for?"

Approach 2: Upload your data and let AI analyse it directly Tools like ChatGPT Advanced Data Analysis allow you to upload actual files and have the AI run analysis on them.

For users with ChatGPT Plus, this is genuinely powerful: you can upload a CSV and ask questions in plain English.

Natural Language Data Queries

ChatGPT Advanced Data Analysis allows you to query data without writing formulas:

• "Which sales rep had the highest average order value in Q3?"
• "Show me the month-over-month growth rate for each product category"
• "Are there any anomalous data points — months where sales deviated significantly from the trend?"
• "Create a visualisation of revenue by region for the last 12 months"

The AI writes and runs the code; you get the result.

Using AI to Write Formulas and Queries

Even without uploading data, AI is excellent at writing the formulas, SQL, or code you need:

"Write an Excel formula that looks up a customer ID in column A, finds their most recent purchase date in a separate table, and returns a status based on recency: current if within 90 days, at risk if 90-180 days, and churned if over 180 days."

This works for Excel and Google Sheets formulas, SQL queries, Python pandas operations, and R data manipulation.

Interpreting Results

Once you have data results, AI can help with interpretation:

"Here are the results of my analysis: [PASTE DATA]. The North region shows a 15% decline while all other regions grew. What are the most likely explanations for a region-specific decline? What data would help me confirm or rule out each explanation?"

Important Caveats for Data Work

AI makes calculation errors Even with Advanced Data Analysis, verify important calculations independently.

Correlation vs. causation AI will identify patterns in data. It may suggest causal explanations. Treat causal claims as hypotheses to test, not conclusions.

Data privacy Be careful about uploading sensitive data to consumer AI products. Review your organisation's data policies before uploading customer data, financial records, or anything confidential.