This week: Ai Doing Your Data | Brawn Documentary Trailer​

---

Hello Reader,

Can AI analyse your driving data?

I mean, can it actually, in any useful way, take your data, analyse it, and then make accurate and useful suggestions on how you could improve your driving?

Seriously, could it?... and just by "looking" at it.

Well, just maybe it can... Read on 👇

Data is harder than it should be

If you’ve ever tried motorsports data analysis, you'll understand that despite its potential, it's generally pretty tricky to extract as much value from it as you’d hope. If you're not using this tech every day—or even if you are—the user experience can be frankly terrible. It is archaic, slow, and, much of the time, overwhelming.

Despite the wealth of help available 👋, many purchase a data logger with all the best intentions, then, after a few valiant attempts, never look at the numbers again. Without assistance, it’s rare to delve into your data at the track. If you're keen, then perhaps, with mixed emotions, you might review it back home on Monday, thinking, "Oh, if only I’d known that at the weekend..." or is that just me? Haha. 😆

Ai promises the world. Can it help you - today?

Ai is all the rage with amazing promises for a future we can hardly imagine.

But, is Ai any good to you, today?

Well, yes, I think maybe - if you know how to "talk" to it.

The Excel Bat Phone

I used to offer a kind of Excel support line, "the Bat Phone," for my Olympic sports clients. It wasn’t that they couldn't use Excel —everyone is an expert, right? 🤣 — but they would get stuck, a lot.

When they got stuck these highly paid people would then waste hours - and hours - searching Google trying to find what was often "such a simple answer."

So instead we set up a service where they could call me. I've some Excel form and understood what they were trying to do. So instead of Google, I'd provide them a solution, and much quicker.

The point here is that I didn't always have their solution at the forefront of my mind either, but I could find it much more quickly. My "Google-fu" was strong and so it worked out really well.

In the last year, I've been developing what I call "Ai-fu", but what most people are calling "prompt engineering." In the same way, the idea is to get the ai systems to work more effectively for you.

Guess what? I’m getting pretty good at it...

...and then ChatGPT added the ability to upload images.

I mean, could it???

Lap Time Finder

You might know I've been working on a tool called "Lap Time Finder." Progress hasn’t been super fast, mostly due to tyres (because what in motorsports isn’t about tyres?) and some early roadblocks with writing code and accessing data.

But What If...

What if instead of uploading data files my system can't read, you simply pulled up a few traces, took a picture of them, gave them to a chatbot - with a few well-crafted prompts - and see what it made of the data?

Surely it couldn’t be that insightful? Could it?

I mean there’s NO WAY it could identify:

• the top three opportunities on the lap to improve,
• offer a potential lap time gain for each opportunity,
• pose questions ask the driver, or
• assess the driver’s skill level to make tailored suggestions for improvement.

Not from one picture.

Not accurately.

No chance.

Well, in this weeks article you can see for yourself:

Click to see How Ai Can Analyse Your Data - Today

​

And here's a sharable link if you want to canvas your friends opinion in this too:

​https://www.yourdatadriven.com/ai-accurately-analyse-your-motorsport-data/​

I encourage you, as always, to give me feedback and thoughts if you try my advice.

Right now, I encourage you to:

• fire up your laptop,
• pull up some recent data,
• get the data channels I’ve detailed on screen (for the GPS-corrected version, read this, but its non-essential)
• take a picture, and
• submit it to ChatGPT 4 with my prompts.
• See what comes back.... 🤯

Then let me know how you get on!

The future is here I think?

Good luck staying ahead of the curve,

Samir

​

p.s. Brands Hatch, used in my example is a super simple track. I'd be really interested in whether the ai actually works for you on something more complex. How pleased are you are with the results?

For context, in my example, my driver is in his third season and benefiting from excellent coaching, achieved his first podium at this event, leading much of the race and narrowly missing a first win due to an unexpectedly extended race. He end up consistently clocked 54s (per ai suggestion...), with potential for 53s (which would be a class lap record.) So he's on the brink of joining the front runners (per ai analysis) and this advice would have helped him I'm sure. Fascinating and terrifying all at the same time!

​

p.p.s. Quick heads up! Whilst on the subject of data, I'm planning a very special data themed "Black Friday / Cyber Monday" offer around my data fundamentals course. The off season is the perfect time to up-skill yourself so if you were thinking of getting this, hold off until Nov 24th, so you can benefit from the deal. More info soon.

​

Fun from around the web:

• The Brawn documentary launches Nov 17th. If you've not seen it, here's the trailer:

​

​