← All courses
AI Data Analyst
Course · Data and analysis · Level 2
Turning numbers you have stared at for years into things you had never seen, without becoming a data scientist.
Ask the fresh eyes question and spot patterns that sat under your nose for years.
A messy export gets cleaned up (broken encoding, mixed date formats, duplicate rows) while you make coffee.
Cross sales against warehouse stock, and the real bottleneck shows itself.
Key figures come out of a PDF report, and you know how to check they are the right ones.
Build a dashboard that answers "so what do I do?" before anyone asks it.
One script stays with you and runs again next month, with no code written by hand.
They stand alone but read best in order.
Adam has looked at the same numbers for five years and stopped seeing them: that is familiarity blindness. The first question that brings a file you know by heart back into focus.
Broken encoding, three date formats, duplicate rows. Upload the file and take it back clean, with a list of everything that was fixed.
Stop ordering the usual house red. Describe your dataset and get analyses outside your repertoire, each with the reason it makes sense.
Rahul opens three systems to answer one question. Join sales and stock on the right key, and the bottleneck surfaces on its own.
Half a day of reading faces David, plus an investor report dozens of pages long. He leaves with the numbers in a table and a way to verify them.
Twenty-seven slides go to the CFO, who stops at three. Learn to answer in one page the only three questions a decision maker asks.
Every month Sofia rebuilds the forecast from nothing. Write the recipe once: swap the file, the result adapts, and the work survives your holidays.
All seven parts run on the course datasets, with the prompts already written.
Download them from the lesson. They are genuinely messy, up to the hundred and eighty-five thousand row file in the last part.
Use ChatGPT or Claude. A free plan carries you to the dashboards part. The last part needs a subscription that includes Claude Code.
Stuck? Ask, in writing or out loud. It answers about the lesson you are on.
5 questions per lesson. At the end you get a course certificate, with the quizzes you passed.
You ship reports every month and you know those files hold more, but you have no idea where to start digging.
You decide on numbers spread across different systems, and would rather see them all at once.
Comfortable in a spreadsheet is enough. Bring a file from your own work and you can try everything on real numbers.
DigComp is the list Europe uses to name digital skills. There are 21 in all: this course covers 9. Below you find which ones, and in which lesson you learn them.
For data leaving the company: Your Data's Journey with AI.
Vibe Coding comes next.
For several people, team licences cost less per person.
If after 10 minutes you have not learned something new, you have lost 10 minutes.
No. Not a single line gets written by hand: you describe what you want in your own language, the AI writes the code, and you read and correct it the way you would a spreadsheet formula.
Three hours of lessons, nine if you redo every exercise on your own files.
It depends on your plan, as part 1 explains. On personal plans conversations can be used to train models. On work plans they cannot. The course datasets are synthetic, so practice runs there.
Yes. A whole part works on an export with broken characters and US date formats.
Max Turazzini. 3 years of workshops in more than 200 companies, 2,000 people trained. It is built for companies that have plenty of data and no analyst to ask what it says.
"You learn AI by doing it, by getting it wrong, and by having someone beside you who has been there before."
Start with the free lesson →