Data Analysis with Claude

You stop doing the analysis and start directing it.

Claude can now read a spreadsheet, answer questions in plain English and build a report in minutes. It can also be wrong, and sound just as sure. The job has not changed: ask a good question, check the answer, and show people something they can act on. That is what I teach.

The tools are changing. The fundamentals are not.

Start the five-day challenge See four of the cards first

Twenty cards. One idea each.

These are the habits I teach on my live course, one per card. Each card gives you the idea in a few lines, a prompt to paste into Claude, a quick check on the reply, and one thing to try on your own data. Four of them are below. The full deck of twenty is a free PDF. Download it and keep it open beside Claude. You do not need to give your email address.

One idea. One prompt. One check.

Before you askCard 1 of 20

Check what is in the file before asking questions

Why this matters

Before asking Claude to analyse your data, ask what it can see in the file. Claude will list the sheet names and column headings. This also helps you quickly check that you attached the correct file.

The prompt

I have attached [file name]. For each sheet, list the sheet name and the column headings. Do not analyse anything yet.

Check the result

Compare Claude's list with the file. If a sheet is missing or a column name looks wrong, stop and check why before continuing.

Try it with your data

Try this with the spreadsheet you use most. Copy the column names exactly as Claude shows them, because you will need them in the prompts that follow.

Check the resultsCard 5 of 20

Sounding confident does not mean the answer is correct

Why this matters

Claude can sound confident even when the answer is wrong. It may estimate instead of calculate, or use the wrong column. Before trusting an important result, check one simple number yourself.

The prompt

How many rows of data are in the [Sales] sheet? Give the number and say exactly how you counted it.

Check the result

Open the file and go to the last row. Subtract one from the row number to allow for the heading row. Compare this number with Claude's count. If they match, you can trust the next results more. If they do not, stop and find out why before trusting anything else in the chat.

Try it with your data

Choose one figure in your file that you can check in ten seconds. Make it the first thing you ask for in every new chat.

Better questionsCard 12 of 20

Ask Claude to suggest useful questions

Why this matters

Most prompts ask Claude a question. This prompt asks Claude to suggest questions you could explore. It works with any file. Treat the answers as ideas to investigate, not as proven facts.

The prompt

Explore this data and tell me three things worth knowing. Show the evidence for each one.

Check the result

Choose the finding that surprises you most. Ask one more question to check whether it is true. Until you check it, treat it only as an idea to explore.

Try it with your data

Start with the open question. Then try more specific questions, such as three risks, three things a new manager should know, or three things that have changed since last year.

Messy dataCard 17 of 20

Find the problems before fixing them

Why this matters

If Claude fixes everything at once, it is difficult to check the changes. Start by asking it to list the problems without changing anything. Then compare its list with the problems you noticed during a quick check of the file.

The prompt

I have attached [export file]. Do not fix anything yet. Check the [Sales Data] sheet and list every data quality problem you can find, with the number of rows affected by each. Check for duplicate rows, blank rows, numbers stored as text and inconsistent spellings. Show the findings as a table.

Check the result

Check for problems that Claude may have missed. For example, does the total always equal quantity multiplied by price? Looking through the file yourself and using Claude's list will find more problems than either method alone.

Try it with your data

Choose a file you do not trust. Look through it first and write down any problems you see. Then run the prompt and compare Claude's list with your own.

Download all twenty cards (PDF)

23 pages. Print them or keep them on screen.

Spreadsheet to Answer: Five Days with AI

You stop doing the analysis and start directing it.
Five days. Twenty minutes a day. Then your own data.

Each day answers one question you have been asked at work. You watch a short video and follow it on a sample sales file. Then you try the same step on your own spreadsheet. Every day ends with the check, so you know the answer is right.

Day 1: What is actually in this spreadsheet?
Day 2: Why will this export not pivot?
Day 3: What were sales by region, and is that right?
Day 4: What changed, and where?
Day 5: Can I see that as something I can click through?

Start today

Sign up on the challenge page. Day 1 arrives in your inbox within a few minutes. Then one lesson arrives each day for the next four days.

Each day has a short video, a file of prompts to download, and a check so you know the answer is right.

After the challenge, I send one short email a week with a prompt card and a worked example. You can leave at any time.

Start the five-day challenge

The challenge runs on my sign-up page, hosted by Systeme.io. Joining is your choice, and you tick a box to confirm it. I only use your address for these emails. To leave, use the link in any email or write to ian@ianlittlejohn.com. The privacy policy explains how your details are handled.

What I teach

Good analysis has always meant the same four things, whatever the tool. Each one is a habit you can practise.

Clean: make the data trustworthy first
Structure: know what is in the data before you ask
Analyse: metrics, comparisons, trends, rankings
Communicate: results people can act on

I teach with Claude because it is the fastest way to turn a plain-English question into an answer you can check. The habits carry over to whatever tool you use next.

Live course

Data Analysis with Claude

A live, small-group course in four sessions. You start by attaching your first file. By the end you are cleaning a messy export, reading hundreds of customer reviews at once and pulling in data from the web. Every section ends with an exercise on your own data, so the skills leave with you.

It is for people who work in Excel and are handed data to make sense of: analysts, finance and operations managers, and anyone expected to have an answer by the end of the day.

Dates for the next course go to the list first.

Who is teaching

I have spent more than twenty years teaching business professionals to work with their data, and my Excel courses on Udemy have reached hundreds of thousands of learners. My approach is the same every time: why it matters at work, then clear steps, then practice on your own files. Data Analysis with Claude is that approach on a new tool, and it is the most useful thing I have taught in years.

- Ian Littlejohn

More about Ian

Three questions people ask

Do I need a paid Claude plan?

Yes. You need a paid Claude plan, Pro or above, and the Claude desktop app for Windows or Mac. The free plan does not have the file and analysis features the cards and the challenge use. If you use Claude at work, check your employer's rules before you upload anything.

Is this for developers?

No. It is for people who work in Excel and are handed data to make sense of. There is no code to write. The skill is asking precisely and checking carefully.

How often will you email me?

Once a day for the five days of the challenge. After that, at most once a week: a prompt card and a worked example, or course dates. Every email has a way to leave.

Five days. Then your own data.

Start the challenge today. Day 1 is in your inbox within a few minutes.

Start the five-day challenge