How to learn Python in Excel from Stringfest Analytics

So you, an Excel user, are interested in learning a bit of Python. Where do I suggest you start?

Python in Excel changed the honest answer to this question. It used to be install Anaconda, set up an environment, learn the language, and good luck. Now you can write your first useful line of Python in a workbook you already have open.

Here is the ladder I sketch out for my audience. Every rung says what it is, what it costs, and the signs that you are ready for the next one.

You do not have to buy anything from me to get started, which is why the free rungs go first. And if you jump ahead to a paid rung and it does not stick, come back down one.

1. Read a few posts, sit in on a free session

What it costs: nothing, and about an afternoon.

Before you spend a dollar, get one result. Work through a few of the Python in Excel posts on the blog, most of which come with an exercise file, and sit in on a free live session from the events page.

The point of this rung is simply finding out whether Python solves a problem you actually have. Plenty of people run through a couple of posts and realize their real bottleneck is Power Query, or a PivotTable, or a source file nobody has cleaned up since 2019. That is a useful thing to learn for free.

Move up when: you have seen Python do something you know your formulas cannot do, and you want to do it yourself.

2. Use LinkedIn Learning through your library

What it costs: very often nothing, and a few hours per course.

This is the rung so many have no idea about, so I put it near the top. LinkedIn Learning comes free with a public library card in Ohio and in a lot of other library systems. Check with your local library before you pay for anything, mine included. If your card works, the whole catalog opens up: Excel, Python, analytics, all of it, for the price of a library card you already have.

My course Data Visualization with Python in Excel lives there too. It is the most specialized thing on this page, so bookmark it and come back to it after rung 4. Charting is where Python in Excel pays off most visibly, and it goes a lot better once you know how data gets into a PY() cell in the first place.

Move up when: you can follow along in a video and then stall the second you point the same technique at your own workbook.

3. Read Advancing into Analytics

What it costs: the price of a book, and a few weeks at a comfortable pace.

Advancing into Analytics: From Excel to Python and R (O’Reilly) runs alongside this ladder more than it sits on it. It came out before Python in Excel existed, so it teaches the wider Anaconda world rather than PY() cells.

That is exactly why it holds up. It gives you the foundations underneath the feature, so you know what a DataFrame is and why anybody bothered inventing one. If you only ever want quick wins inside a workbook, skip this rung. If you want to know what you are doing when something breaks, do not.

One practical note so you do not buy it twice: the ebook is included free with Python for Excel in a Week down at rung 6. If you already know you are headed there, wait.

Move up when: the concepts feel comfortable and you want them working inside Excel specifically.

4. Take the Python in Excel Crash Course

What it costs: $19, and under an hour of video.

This is the mechanics rung. It covers the initialization environment and the packages that load with it, how objects behave so you can save and reuse them, and the several ways of getting data in through cells, tables, and Power Query. You build a dynamic moving average chart as a first real win, and you see where Python fits next to Advanced Analysis and Copilot. Exercise files and links to the Microsoft documentation come with it.

Almost everybody who gets stuck in Python in Excel is stuck on one of these mechanics rather than on Python itself.

Move up when: you can get your own data into a PY() cell and get something back without looking it up.

5. Work through Python in Excel: Quick Wins

What it costs: $15, and fifteen short videos.

Quick Wins is a reference library. Fifteen videos, one use case each: random sampling, missing values, descriptive statistics, correlation matrices, frequency tables, resampling for time series, indexing by group, leading and lagging variables, moving and cumulative aggregations, and a good run of charts from pairplots and jitterplots through small multiples, jointplots, and bubble plots. A cheatsheet comes with it.

A reference is worth a lot more once you know what you are looking at. Take it after the mechanics have clicked, keep the cheatsheet open, and fold one use case into a real workbook of yours each week. Your own GL, your own AR aging, whatever is already sitting on your desktop.

Move up when: you have watched plenty and built little, or your own data keeps breaking what the videos taught you.

6. Come to a live session

What it costs: see the current listing, and either one afternoon or five lunch hours.

A recorded course cannot answer a question about your workbook. That is the whole reason this rung exists. Everything is recorded anyway, so you are not giving up the replay by showing up live.

There are two, and they are not the same thing.

Dates move around, so check the events page for what is open right now. If nothing is running when you read this, get on the list there and you will hear when the next one opens.

Move up when: the skill is yours and the problem changes from learning this to keeping up with it.

7. Join the Modern Excel AI Stack membership

What it costs: monthly, and as much or as little time as you want to give it.

This field moves faster than any one course can keep up with. Python in Excel picks up new packages, Copilot picks up new tricks, and something is different every quarter. Nobody wants to start over each time.

The membership is where the ongoing content and assets live, and it includes a question drop, so you can put your own question in front of me instead of guessing. This is the rung for people who already have the skill and now want to stay current.

How the rungs fit together

These are not seven separate things to buy. They answer different questions, in order.

• Rungs 1 and 2 cost nothing and tell you whether you want this at all.
• Rung 3 gives you the concepts underneath, and it is optional.
• Rung 4 teaches the mechanics of the environment.
• Rung 5 hands you the use cases to apply them to.
• Rung 6 is where somebody answers your question out loud, about your data.
• Rung 7 keeps you current after the skill is yours.

Most people do not climb all seven, and that is fine. Plenty of Excel analysts get everything they need out of rungs 1, 2, and 4 and never look back.

If this is for your team

The ladder above does not transfer cleanly to a department. One person picking their way up seven rungs on their own time is a different problem from getting twelve analysts to the same place by Q3. A few things I would say to any manager sizing this up.

Start with the work, not the feature list

The usual instinct is to list what people should learn. Pandas, plotting, forecasting, maybe Copilot while we are at it. The more useful question is where the team is losing time right now. Who rebuilds the same monthly report from scratch. Who is copying data between workbooks by hand every close. Which one person has quietly become the department’s Excel help desk because every weird workbook eventually lands on that desk.

Those answers scope training far better than a syllabus does, and they tell you which rungs your team can honestly skip.

Sort out the environment before you schedule anything

Python in Excel needs a license that includes it and a Microsoft 365 channel that supports it. Locked-down corporate environments tend to have opinions about both, and so do virtual desktops, bring-your-own-device shops, and regional IT teams reading the same policy three different ways.

I have started enough sessions where half the room could not run the first exercise to say this plainly: run the tech check weeks ahead, not on the morning of day one. I wrote up how I run one, with free downloads, and you are welcome to use it whether or not you ever hire me.

Buying seats is not the same as training a team

You can put a group through the on-demand courses at rungs 4 and 5, and for a genuinely self-directed team that works fine. Just know what you are buying. Self-paced material has a completion problem, and the people who most need it are usually the ones least likely to finish it on their own time.

If it matters to you that the whole team actually arrives somewhere, put something live on the calendar so showing up is part of the job rather than a good intention.

Aim at judgment, not a feature tour

The point of team training is that fewer wrong decisions happen afterward. People spend less time overbuilding, less time unwinding brittle logic, and less time defending analysis that does not hold up. That comes from familiar examples, real business data, and enough practice time that the team can handle the next messy export rather than only the one file we walked through in class.

If that is the kind of thing your team needs, my How I work page explains how engagements get scoped, or you can just get in touch and we will figure out whether it is a fit.