From PDFs to Pit Lane: Building a Real-Time Data Product for McLaren Racing

In this week’s episode of the PolicyViz Podcast, I chat with Michael Gethers, former Head of Data & Strategy for the McLaren IndyCar team, about how a personal side project analyzing IndyCar timing PDFs turned into a job building real-time data tools for a professional race team. We dig into what it’s like to design data products for engineers, strategists, and drivers who need to understand information instantly while a car is on track. Michael shares how he moved from making public visualizations on Twitter to building an internal analytics application from scratch, why “pretty charts” weren’t enough for the engineers, and how user feedback shaped the product. We also talk about race strategy as a probabilistic data science problem, the difference between dashboards and data products, and what he learned about designing for cognition under extreme time pressure. If you care about dashboards, data storytelling, or building tools people truly use, this conversation is a goldmine.

Resources

Guest Bio

Michael is a data scientist with over a decade of experience spanning the full data stack, from building pipelines and infrastructure to crafting beespoke visualization applications.  A UC Berkeley Statistics graduate, he has held data science roles at Salesforce, HUMAN Security, and McLaren Racing, where he led the IndyCar team’s data function and served as a race strategist.  He currently resides in Indianapolis, IN.

Different Ways to Listen to the Show!

New Ways to Support the Show!

With more than 250 guests and 10 seasons of episodes, the PolicyViz Podcast is one of the longest-running data visualization podcasts around. You can support the show by downloading and listening, following the work of my guests, and sharing the show with your networks. If you’re interested in financially supporting the show, you can sign up for my Patreon platform, make a one-time payment via PayPal, or shop one of the show’s sponsors.

Transcript

00:01.14
Jon
All right, there we go Hey, Michael, good to good to see you. Good to meet you. Thanks for coming on the show.

00:05.82
Michael
Yeah, of course. i’m Glad to be here, John.

00:08.06
Jon
um Okay, so I’m pretty excited to chat with you for, i would say two reasons. First, I learned of your work through the Dashboards That Deliver book.

00:19.54
Jon
um So I’m excited to talk to you just generally about dashboarding and creating things. um Also because working on a race team, like the amount of data you must have seen is just like makes my mouth water a little bit.

00:32.83
Jon
But also my son is very much into racing, wants to get into the field. And you’ve already connected us with some other engineers. So thank you for that in advance. But so so a couple of different reasons here.

00:43.47
Jon
um So I thought maybe we’d start with a little background. And um you know how did you you you you were the lead of data and strategy at McLaren.

00:48.36
Michael
Thank you.

00:52.91
Jon
And so maybe we just start like how you got to that spot and then what your role was what what was there.

00:58.90
Michael
Yeah, that sounds great. um I think I took a fairly unconventional path probably to get to motorsport. um A fortunate one at the end of the day.

01:09.72
Michael
um But sorry, we might have to cut this out.

01:15.73
Jon
no No, it’s good. but we can We can start over. No, its it all gives a little bit. So, ah yeah, so we can start over. But um i did I did an interview once with Lane Harrison, who I don’t remember if he was at where he is now, but we did it outside Visualize Conference at the aquaatorium as it aquarium, I think, in Boston.

01:35.98
Jon
And it’s literally we’re outside and you can hear ships passing by and stuff like that.

01:39.16
Michael
yeah

01:39.50
Jon
So, yeah, so it’s fine. It’s fine.

01:41.14
Michael
So you’ve had you’ve had worse then is what you’re saying.

01:41.18
Jon
Not a problem. i’ve had i’ve had way I’ve had way worse and way worse audio quality.

01:43.64
Michael
Yeah. Okay.

01:45.58
Jon
Yeah, not a problem.

01:45.69
Michael
Yeah. I mean, i invent I invested in this microphone to try to limit the sound, but I can just hear it. It’s right there.

01:51.22
Jon
That’s great. How old owl is he?

01:51.70
Michael
But yeah, he’s two.

01:52.92
Jon
He? She?

01:54.04
Michael
He’s two.

01:54.17
Jon
Two. Okay.

01:54.52
Michael
Yeah.

01:54.73
Jon
Yeah.

01:54.92
Michael
We’ve actually got another one due in two weeks.

01:55.05
Jon
Okay.

01:57.92
Michael
So well yeah, we’re’re we’re down to it here.

01:58.42
Jon
Oh, wow. Okay. we definitely have to get We definitely have to get through this then.

02:02.20
Michael
Yeah, exactly.

02:02.28
Jon
So um that’s okay.

02:02.92
Michael
Yeah. um

02:04.40
Jon
Why don’t we just pause and then you can just start your answer and we’ll do it’s fine.

02:07.00
Michael
Sounds good.

02:07.56
Jon
It’s not a big deal.

02:08.66
Michael
Okay. Yeah. So I i came into motorsport, I think from a fairly unconventional unconventional path. um And ultimately, ah you know I’m very grateful for the for the opportunity that I had and the the lanes that opened up for me. But my my background really is is as a core data scientist. I studied statistics at at UC Berkeley.

02:31.54
Michael
um I worked in the Bay area did the tech thing in the Bay Area, i worked as an actual, actually my first job out of out of college was as an actuarial analyst. And then i kind of um worked at the a number of different tech companies, most notably in San Francisco was Salesforce.

02:40.62
Jon
Oh, all

02:48.58
Michael
And then um pivoted to some cybersecurity work after that. Um, but always, you know, since I was a kid, you know, I i went to the racetrack with my dad, specifically the indian Indianapolis motor speedway with my dad, um, since I was five years old.

03:04.38
Michael
Um, you so he, he took me and I just loved it.

03:04.74
Jon
right.

03:08.15
Michael
You know, it was, it was something that I grew up with and it was sort of a bonding thing between me and him.

03:08.70
Jon
Yeah.

03:12.43
Michael
as we were, as I was growing up. And I went to my first race when I was 10 years old. And I’ve been to every Indy 500 since then, um except for the COVID year.

03:22.33
Jon
Wow.

03:25.09
Michael
um

03:25.46
Jon
Right.

03:26.07
Michael
So it was always kind of a part of who I was.

03:26.17
Jon
Right. Yeah.

03:28.95
Michael
and um Because of that, actually, I just I and being a data nerd, as many of us are, and many of the people probably listening to this podcast are, um I just found that I was

03:37.29
Jon
Yeah. Yeah. yeah

03:44.79
Michael
kind of drawn to finding more data about the race because racing is a sport where, you know, it’s not like basketball or or football really where everything you see, everything that’s happening, you can see in front of you, right?

03:49.06
Jon
Mm-hmm.

03:59.42
Michael
Um, if you’re watching a race on TV, you can see a couple of cars at a time, or maybe a couple of corners at a time. um even if you’re at the race,

04:10.65
Michael
the tracks are massive, right? You just can’t see the whole thing. um And so I remember one year being at the Indy 500 with my dad and um and I just had questions about what was happening, what was going on um during the during the race that I just couldn’t answer.

04:27.81
Michael
And there was no way to get that that information at the time. and so And so what I did after the fact was I

04:31.49
Jon
Right.

04:36.70
Michael
I just kind of scoured the internet for for data. And actually IndyCar, interestingly enough, was ah they published public PDFs that contain some timing data, not extremely robust and not extremely easy to use, but they published some of that timing data.

04:51.29
Jon
Yeah.

04:53.30
Michael
And so I just scraped that and started doing some analysis and I put it on the internet.

04:58.11
Jon
Mm hmm.

05:00.60
Michael
um

05:01.11
Jon
Yeah.

05:01.62
Michael
So I created it. Well, I just had a Twitter account for myself originally.

05:06.26
Jon
Yeah.

05:06.26
Michael
and started posting things there. i Through that, I actually got to know um um one driver in particular, who was J.R. Hildebrand. um And he um was just interested in what I was doing.

05:21.55
Michael
He happened to be sponsored by Salesforce at the time. um

05:24.66
Jon
Oh, interesting.

05:25.75
Michael
So there was that kind of natural connection because I was working at Salesforce when I was doing this. And that was kind of my first just like in a little bit just because you know he let he invited me to come to the track as part of the Salesforce demo, basically, that they were doing at the track.

05:35.58
Jon
Yeah.

05:42.19
Jon
Oh.

05:43.44
Michael
And that was kind of my first, just a little wedge in.

05:46.63
Jon
Yeah.

05:47.67
Michael
And then fast forward, you know that that was kind of a one-time thing. um Fast forward a ah few years and um I had just left my cybersecurity job. My wife and I were living in London.

05:59.09
Michael
We were moving from London, we’re planning to come back to the States by way of Switzerland, which is a longer story that i we don’t need to get into here. but um

06:07.19
Jon
yeah

06:08.92
Michael
When I was in Switzerland, I was like, you know what? I really love doing that IndyCar analytics thing. So I’m going to start doing that again.

06:15.48
Jon
Yeah.

06:17.43
Michael
um But this time I’m going to do it bigger and better. right And I wanted to treat it as a learning experience too.

06:20.59
Jon
Yeah.

06:22.30
Michael
I’d always wanted to learn D3 because I’d never i had never known really anything about front-end web development. And it was just kind of something I had always been interested in. So um i created a new Twitter account.

06:36.15
Michael
um It was called Rose of Three, um which anybody who’s familiar with the Indianapolis Motor motors Speedway, the Indianapolis 500, the cars start in 11 Rose of Three.

06:38.42
Jon
yeah

06:46.88
Michael
That’s where that comes from. So my my Twitter ah handle was Rose of Three. I created ah an accompanying website called Rose of Three dot com. And I just kind of went all in like it was my full time job, basically creating creating analytics content basically um around IndyCar specifically.

07:03.16
Jon
h

07:05.14
Michael
And it was easy because um I just had so many questions.

07:06.08
Jon
Right.

07:10.69
Michael
And so when you had so many questions, you just want to kind of chase after it and answer it, answer those questions.

07:11.42
Jon
Yeah.

07:15.37
Michael
And, um you know, I, it’s, it’s kind of funny for me to look back at the evolution of, of my, of my, my work there because I could see myself getting better over a period of months just because you get better at asking the right questions.

07:26.14
Jon
Right.

07:28.30
Michael
You get better at the actual technical development part of it.

07:28.82
Jon
Yeah.

07:31.42
Michael
um you know My website, I look back at the website that I made now and it’s um you know it’s clear that it was my first it was my first website, but um I’m still proud of it on some level because it really took me where where i ended up going.

07:37.85
Jon
Yeah. Yeah. ah Yeah. yeah

07:48.54
Jon
yeah

07:49.98
Michael
um So through that, ultimately, Gavin Ward reached out to me, who he was the soon to be team principal at Aero McLaren in Aero McLaren’s IndyCar team, which is the McLaren racing IndyCar team um in Indianapolis.

08:10.20
Michael
Formerly, he um worked at ah Red Bull F1 and he worked at Penske Racing. um in IndyCar as well.

08:20.60
Michael
He reached out and and just asked the question, you know, hey, have you ever have you ever thought about working in motorsport? And I was just kind of like, yeah, yeah, I have.

08:29.82
Jon
Yeah, kind of every day. Yeah, yeah, yeah.

08:31.09
Michael
Yeah. um So sort of that conversation went on for a while to to actually determine what it was going to look like.

08:37.66
Jon
Yeah.

08:39.59
Michael
But a few months later, i ended up joining the team. That was at the end of 2022. Yeah.

08:43.82
Jon
Wow. So let me ask ah a couple of questions. So between when you first played around with those PDF files and then decided to kind of go all in, did the, did the data that Indy was ah publishing, did that change fundamentally?

09:01.46
Michael
Um, no, it did not change but fundamentally, but my access changed fundamentally.

09:03.67
Jon
Okay.

09:08.84
Michael
Um, so a couple things, a couple things happened.

09:08.86
Jon
Okay.

09:12.95
Michael
Um,

09:15.58
Michael
you know i was just kind of I was doing what I could with the with the scraping right of of their PDFs, which um you know they’re publicly posted PDFs.

09:19.54
Jon
Right, right.

09:24.34
Michael
pdf PDF parsing is not ah a precise science.

09:26.70
Jon
Yeah.

09:29.21
Michael
I think it’s better now probably than it was at the time.

09:29.24
Jon
Yeah, I bet.

09:31.17
Michael
But um i did I spent a lot of time doing data correction basically in that process.

09:34.97
Jon
Yeah. be yeah

09:36.31
Michael
um But at one point, I actually reached out to ah somebody. Actually, I can’t remember if i reached out or if they reached out to me. um In any case, somebody ended up giving me better access to to some data that I could play with.

09:49.41
Jon
Gotcha.

09:50.88
Michael
And I was able to get data in in CSV format.

09:51.21
Jon
Gotcha.

09:54.81
Michael
And um that made it obviously much, much easier and more accurate.

09:59.86
Jon
Yeah.

10:00.13
Michael
And obviously, that was not the data that I ended up working with on the team. But that that is what gave me my kind of the ability to do this at ah at a greater scale than I was able to do before.

10:04.63
Jon
Sure.

10:12.14
Jon
I’m curious and you may not know the answer to this, but, um, you know I’ve talked to lots of other people who do similar sorts of things, you know hockey and basketball, other sports where they they they do sort of similar things and then they get access to the to the to more of the data and they have relationships with teams or something like that.

10:29.11
Jon
in Do you have any instinct as to why… they would be, or you would reach out to them but or they would just give you more access to the data. Like, was it, were they at the time they just didn’t have like people doing data viz and data analytics on the, on, on, I mean, there’s a ton of data here and I’m just curious why, like you’re this guy out there just doing this almost for fun.

10:47.03
Michael
Yeah.

10:51.19
Michael
Yeah, I’m just, I’m just, a exactly.

10:51.66
Jon
And they’re like, Hey, hey here’s a bunch of data. Yeah.

10:54.43
Michael
It’s just a guy um

10:55.54
Jon
Yeah.

10:56.34
Michael
You know, I, I don’t know. the answer really i can only kind of hypothesize but i think number one what i was doing was was cool people liked it um it wasn’t it wasn’t just for me it was for a community of people that really loved the sport and um in my mind what i was doing was actually quite good for the sport because it was giving people um

11:09.53
Jon
Yeah.

11:16.34
Jon
Yeah.

11:21.27
Jon
Yeah.

11:23.74
Michael
getting more people engaged in a different way than they were able to be engaged before.

11:27.61
Jon
Right.

11:29.50
Michael
And also you might capture a new audience that they weren’t able to capture before. Maybe there’s the data nerd ah audience that never thought they had a home in racing, but um because because all the data was was so under wraps, right?

11:38.14
Jon
Yeah.

11:43.64
Michael
So i thought, you know, I don’t know what the actual answer is, but I think it’s something, something in that vein.

11:47.13
Jon
Right.

11:49.60
Michael
Right.

11:49.85
Jon
Yeah.

11:50.04
Michael
And, ah yeah, I just sort of had somebody who, who was excited about what I was doing and, wanted help, help me make it better.

11:59.16
Jon
Yeah. I mean, that’s ah that’s what we all need, right?

11:59.83
Michael
Yeah.

12:01.53
Jon
We need those cheerleaders. um

12:03.03
Michael
Yeah. Yeah.

12:03.96
Jon
Okay. So you get to McLaren. um are you now Are you now back in Indianapolis?

12:10.97
Michael
Uh, yeah, I’m,

12:11.06
Jon
like did When you were in London, you moved back?

12:13.34
Michael
Yeah, so you know quick quick history of my my geographyer like my geographical ah um

12:20.31
Jon
Yeah. Yeah.

12:21.82
Michael
evolution. Went to school in the Bay Area, worked in the Bay Area for a while. My now wife and I moved together to London for about five years. um Her family is actually Swiss. And so when we made the decision that we were going to come back to the U.S.,

12:38.42
Michael
and We were like, okay, well, before we do that, let’s just do a year in Switzerland before we come back.

12:42.16
Jon
Okay.

12:42.65
Michael
um

12:42.94
Jon
Okay.

12:43.64
Michael
We were planning to go back to California, which is where both of us um called home before we we left London.

12:49.12
Jon
Right.

12:50.01
Michael
um But then the the opportunity came up in Indianapolis and it it was, you know, I had to take a shot, right?

12:56.42
Jon
Yeah, sure.

12:57.79
Michael
Had to had to had to see what that’s like.

12:57.92
Jon
No, absolutely. Right.

12:59.67
Michael
Yeah.

12:59.80
Jon
Right. Absolutely. Okay. So you’re in Indianapolis, you’re working for McLaren. um What is that day to day like? I mean, where i mean is the McLaren offices, are they…

13:11.74
Jon
I’ve only been to the Speedway once and it was actually last summer for an IMSA race. So I’m not…

13:16.37
Michael
Okay.

13:16.66
Jon
My son’s a big F1 fan and now a GT fan, but he kind of looks down on on NASCAR, especially.

13:20.70
Michael
Gotcha.

13:22.66
Jon
so um So all the racing people who are listening to this could can yell at him.

13:23.40
Michael
Okay.

13:26.61
Jon
um but um are you going like where you’re going to the office every day and like what are you doing with with their data? I mean, are you just ingesting a ton of data all the time?

13:38.23
Michael
Yeah. I mean, it was, it was interesting. Um, yeah looking back at it it’s it’s really interesting now for me i guess um when i when i got there it was almost like

13:46.65
Jon
Yeah.

13:52.60
Michael
they didn’t, we didn’t really know what to, what was the best thing to focus on. Right. They had never had, um, somebody like me, you know, I think on the F1 side, maybe they were, uh, a few years ahead, but not wildly ahead.

13:58.07
Jon
Sure.

14:09.88
Jon
Yeah.

14:09.88
Michael
Um, and I was really the first core data person that was brought into the IndyCar team. And, um,

14:17.46
Jon
yeah

14:19.32
Michael
you know A big part of that was was Gavin’s vision, right? And that’s what we had talked about over a few months before I ended up joining. um So it was it was kind of like, yes, we have all of this data.

14:31.54
Michael
Number one, how is it stored? um number one Number two, like is it queryable? Number three, like what is the actual output that we want from this?

14:36.63
Jon
evening

14:39.71
Michael
What do the engineers need? um

14:41.53
Jon
yeah

14:42.33
Michael
There was a lot of kind of uncertainty around what that was going to look like. um And so it took for me, It was a lot of trial and error, frankly.

14:53.50
Michael
i was trying to Initially, I was trying to put together a lot of the kinds of things that I was i was developing for public consumption on on Twitter and on my website.

15:05.08
Michael
um But I don’t think you’ll be surprised to to learn that that’s not really what the engineers wanted and needed.

15:10.42
Jon
Yeah.

15:11.93
Michael
right um

15:12.76
Jon
Right.

15:13.82
Michael
And I think they initially regarded some of what I was doing as like, oh, this is these are pretty pictures, but where’s the whereas the substance, right?

15:19.13
Jon
Oh, interesting.

15:20.88
Michael
And so it was, fortunately for me, I had i had sort of the the traditional data science background, so I was able to kind of

15:21.50
Jon
Yeah.

15:28.38
Michael
um to kind of meet in the middle there. And and and that’s where i think the application that um that I built and that that I built with the team there ah really kind of took hold a little bit and people started to really get some value out of it.

15:31.80
Jon
Yeah.

15:45.66
Jon
so So were you, i mean, when when when you watch racing or you watch the F1 movie with Brad Pitt, you know, there’s all these screens, people sitting on the on the on the track. um And are you were you primarily focused on sort of real-time analytics?

16:03.38
Jon
Or was it more on the engineering side of, you know, they’re going to test thing? wing, this back wing, this front wing, they’re going to test this other setup on the car and let’s ingest all of the results of those tests and, you know, put it together or, you know, or is it all all, the above?

16:19.61
Michael
Yeah, it was it was probably all of the above, but I think we we tried to we tried to shrink it a little bit at the beginning um to something that’s more palatable, I guess, when you’re starting from from nothing, basically, which we were.

16:22.94
Jon
Yeah.

16:26.42
Jon
the

16:33.72
Jon
Right.

16:34.52
Michael
um

16:34.86
Jon
Right.

16:35.58
Michael
And so what we started with was was timing and scoring data. um And that’s, for those that don’t understand what that term is basically any every race track has a number of a discrete number of timelines around the track.

16:52.95
Michael
And when every car crosses over that timeline, um you know it’s registered. So car seven crossed timeline start finish at X time.

17:00.28
Jon
Yeah.

17:05.93
Michael
right um And it’s a bit more robust than that, but that’s more or less that’s like what it is at its core. um

17:13.59
Jon
Right.

17:14.49
Michael
And there’s a ton of stuff that you can do with just that alone, but it it it requires a lot of kind of careful thought about how you’re going to um how you’re going to transform that data.

17:18.27
Jon
Mm-hmm.

17:26.10
Michael
You can do all all kinds of things, obvious things from like, okay, how fast are we? on every lap, right? That’s very obvious.

17:33.07
Jon
Great.

17:33.61
Michael
How fast are we in each given sector? How fast are we in each given sector if we’re following one car at one second? um How fast are we if we’re in heavy traffic and there’s five cars within within five seconds? um Those kinds of things, you just start to build, you know you start small and just kind of build up.

17:54.07
Michael
that capability. And then once you build up the capability to do the analysis, what we were doing was, okay, how do we present this to people in a way that is is very um digestible very quickly, right?

18:06.47
Michael
Because that’s kind of the name of the game.

18:06.72
Jon
Yeah.

18:08.51
Michael
um The car’s on track, right? And we were building an application to be used while the car was on track in addition to as a kind of data interrogation tool.

18:18.87
Michael
after we’re back in the back in the garage or back in the truck.

18:18.90
Jon
Yeah.

18:22.15
Michael
um And you know in this scenario, in both scenarios, people want data as fast as possible, but especially when the car is on track, it’s kind of like, well, we need to we need to understand exactly what’s going on right now.

18:29.75
Jon
Yeah.

18:35.16
Michael
And was that a good change or was that a bad change? and um ah that That’s kind of where a lot of the sort of optimizations and efficiency gains had to come from that that we ended up building into the tool. um But it’s not just technical efficiencies. It’s also kind of cognitive efficiencies on some level. It’s like, how do you get people to understand this really fast?

18:59.64
Michael
um Which is a ah challenge that i had not I had not really been faced with before.

18:59.70
Jon
Right.

19:04.76
Jon
Right. So, um I mean, I could ask a ton of questions, but um let’s start with the folks. Let’s start with this cognitive question. So you’re so presumably you’re showing, and we could just focus on on on during the race if we want, just to sort of make it easy, but you’re showing visuals to people who may not have seen them in this way before. Maybe they’re way more visual than tabular maybe is how they were used to them.

19:30.82
Jon
So, you know, did you sit down with folks? Like, did you do kind of like formal or informal, like user testing to see what would be the most useful, intuitive, immediate visual that they could get?

19:42.68
Jon
Like, how did you work with the team to like build something that that you know, that you knew that they could use?

19:43.86
Michael
yeah

19:48.73
Michael
Yeah, that’s a great question because initially I didn’t and i initially I was just like, I’m going to go build what I think, what I think we need, which is not the right way to go about it.

19:54.07
Jon
Yeah, right. Yeah.

19:57.19
Michael
Um, obviously, but you know, you live and you learn.

20:00.54
Jon
Yeah. Right.

20:00.79
Michael
Um, yeah, for, for me, it was kind of finding the, the

20:01.06
Jon
yeah

20:07.83
Michael
i hesitate to champions of the product that we were building but but the people the power users who i thought were going to use it the most um and understanding what they really what their gaps were in the tooling that they already had they’re obviously they’re you know extremely talented engineers themselves right they’re not unfamiliar with looking at data right this is what their entire correct yeah yeah so they’re you know they know what

20:13.31
Jon
Yeah.

20:28.41
Jon
Yeah, I mean, right. You’re not dealing with like English lit majors, right? Yeah, right. Yeah. right

20:35.16
Michael
they know the domain 10 times better than I ever will.

20:37.24
Jon
Yeah.

20:39.27
Michael
Right. um

20:40.31
Jon
Yeah.

20:41.56
Michael
But, and they know how to look at data, but they just, you know this is a personal opinion i guess but i think it was sort of proven out as we as we continue to build the tool um the the data tools that had been built that had been built for them before which were all sort of off the shelf products that people were trying to build it was never really done in-house at least not at um the mclaren indycar team um were not I mean, I almost wish I could show them to you. They’re so visually messy and they’re just, there’s so much data on it.

21:19.22
Jon
Yeah.

21:19.48
Michael
that um it’s you know those engineers are extremely adept at making heads and tails of it. um

21:28.09
Jon
Right.

21:28.66
Michael
But for this kind of quick hit, give me a, you know give me where are we fast, where are we slow compared to x y or Z team or driver or you know in this kind of scenario, it was hard to it was hard to parse, right?

21:38.52
Jon
Yeah.

21:43.93
Jon
Yeah.

21:44.09
Michael
um And so that’s where, you know through those conversations and through just observation really, because and you’re you’re talking about the the the Brad Pitt movie, the F1 movie and any other sort of popular culture representation of what motorsport is, it it looks very high tech and then you see the big screens and people on the timing stands.

21:56.18
Jon
Yeah.

22:04.98
Michael
you know I was very fortunate. i I was one of those people, right? I got to be, i was on the timing stand in the pits.

22:08.94
Jon
Yeah. Yeah.

22:11.46
Michael
um And I could look at my screen and I could look at their screen. I could see what they were looking at. I was able to talk to all the engineers, the drivers, um as we were kind of in a practice session and in in races and in qualifying.

22:27.42
Michael
um So I had that feedback loop and the feedback loop was kind of built into the job, which was which was great. right I don’t think I could have built what I built without having that opportunity.

22:35.86
Jon
yeah

22:38.50
Michael
Like if you were just a um you know if if if If you didn’t actually have the exposure to what that world is like and what those needs are, it’s quite…

22:51.22
Michael
you know It’s an esoteric domain, actually. Like it’s it’s it’s complex and it’s um it really doesn’t apply anywhere else in the real world.

22:54.78
Jon
Sure.

23:00.57
Jon
Yeah.

23:01.62
Michael
So, um well, people argue with that probably, but it’s just a very unique domain.

23:08.09
Jon
Yeah, for sure.

23:08.25
Michael
um And it’s it’s difficult to understand what it’s like unless you’ve seen it. And I was fortunate that I i had the opportunity to see it and experience it. And I i think that helped me build things better.

23:17.88
Jon
Yeah, no, for sure. um so So my first question is, when you are in the timing stand ah during a you know during a race weekend, um are you are you are you what what are you doing ah during those times? Are you adjusting the view? are you like what what is And what can you do quickly enough that will be useful to the rest of the team?

23:43.03
Michael
Yeah. um Well, you know, my role on the team, you know, everybody kind of wears multiple hats. These teams are not are not huge. At least the IndyCar teams are are not are not huge.

23:52.95
Jon
Right.

23:57.12
Michael
The F1 team is much bigger. um But. you know, my role on the timing stand was as a race strategist, which was a, a a difficult role to be thrown into having never been a race strategist before or having been around, around that.

24:11.32
Jon
Right. Right. Yeah.

24:12.97
Michael
But, um, At the same time, it kind of makes a lot of sense because a lot of race strategy is is very probabilistic and it’s um it’s kind of, you know what do we think the probability is if we pit now that we will beat this other car out of the pits in five laps?

24:33.45
Jon
Right.

24:34.40
Michael
Or what’s the probability ah right now of there being a yellow flag between now and the end of the race? And how does that influence our our strategy going forward? should we Should we conserve fuel and try to make it to the end hoping we we we get a you know a yellow flag safety car situation? um Or should we just run as fast as we can, splash for fuel?

24:59.03
Michael
And and if that is that the is that the most optimal way to to get to the end of this race? What’s the probability that we’re going to be able to pass any given car on track? Because that’s quite difficult as well, right?

25:09.23
Michael
These aren’t cars running in a vacuum.

25:09.46
Jon
e

25:10.94
Michael
They’re a bunch of cars running on track together.

25:13.18
Jon
Yeah.

25:13.46
Michael
um So this the strategy problem actually does really lend itself to um a data science approach, I think. um And so while was a difficult, you know I felt like getting thrown into the deep end a little bit there, um it was a role that I think, ah yeah, was not ill-fitting on some level, right?

25:35.45
Jon
Yeah, right.

25:36.75
Michael
And so what i’m you know to answer your original question, what what was I doing on stand? it’s multiple things, right? It’s um looking at the data as it comes in looking at the, where our car is relative to other cars on track, keep keeping track of, um you know, who is low on fuel, who’s who’s who’s going to be pitting soon, who looks like they’re able to pass well, who doesn’t. um That’s one side.

26:03.63
Michael
That’s like the very active in the moment kind of thing.

26:05.85
Jon
e

26:06.84
Michael
But there’s also this kind of passive work that’s being done at the same time, which is like, what do I need to build to make this job easier?

26:14.20
Jon
Yeah, easier, right?

26:14.61
Michael
um And so both of those things, I think, were kind of happening at the same time, although the the latter was more observational then than the the former.

26:23.83
Jon
Right, right, right. So, so, so one quick question for you, are there, I, what are the rules in IndyCar about what each team has to share?

26:35.44
Jon
Cause to the point you just made about tracking what other cars are doing, presumably McLaren isn’t collecting data about every car.

26:38.28
Michael
Mm-hmm.

26:44.60
Jon
I mean, maybe they are in Red Bulls, not like there’s a data sharing situation that’s going on and that’s it rules.

26:49.50
Michael
Yeah, so yeah, that it is. Yeah. So um first of all, all of the timing data, everybody can see.

26:56.47
Jon
Yeah.

26:57.72
Michael
So we know not just where our car is around the track.

26:58.06
Jon
OK.

27:01.12
Michael
We know where everybody’s cars are at all times.

27:04.00
Jon
Right.

27:04.63
Michael
um But each team also IndyCar has to share kind of limited telemetry data. um

27:11.18
Jon
Yeah.

27:11.93
Michael
And so that is… I’ll probably get all of them. Maybe I’ll miss one or two, but, uh, speed RPM gear, throttle, brake, um, steering angle.

27:26.01
Michael
Um, and,

27:30.24
Michael
I think there are a couple others. And IndyCar has some unique fields as well. They just switched to hybrids a couple of years ago. So they’ve got hybrid deployment as a field as well.

27:38.22
Jon
Right. OK.

27:42.03
Michael
Anyway, there are a small handful of fields of telemetry signals that everybody has to send to everybody.

27:49.62
Jon
Gotcha.

27:49.87
Michael
So we can see that live actually. It’s at 10 hertz, which is actually lower fidelity than 10 times a second, lower fidelity than obviously what we have on ourselves.

28:00.10
Jon
Yeah. Right.

28:05.97
Michael
But there’s a lot that you can use, that you can do with that.

28:09.66
Jon
and And is that, I have little to no understanding of of the backend side, but is that part, is is ingesting that data, I assume ingesting that data and getting it as quickly and efficiently to the timing stand or to the race director or whoever, that’s on your team. That’s your team’s responsibility to, like they just give it to you and then you’re you’re figuring out how to get it in quickly and and and efficiently.

28:36.79
Michael
Yeah, that’s right. And um yeah, so IndyCar bears no responsibility for what we do with the data that they send.

28:37.91
Jon
Yeah.

28:43.42
Michael
They just promise to send it in some form or another, basically.

28:44.09
Jon
Yeah. Right. Hopefully not as PDFs, but yeah.

28:48.60
Michael
Right, right. um So although i you know I have experience with the parsing of PDFs, so…

28:53.94
Jon
Yeah, right. and that’s That’s right. You’d have a big leg up if they were just giving out PDFs.

28:55.86
Michael
Yeah, yeah, yeah.

28:57.26
Jon
Yeah. but

28:58.12
Michael
um No, so it was that was sort of all on us to to build the infrastructure around that. And, you know, when I joined the team, um All of that data was kind of just stored in log files on a server someplace.

29:13.66
Michael
And you know I mentioned queryability, like none of it was queryable. You had to kind of know what you were looking for and find the right session at the right you know in in the right directory um to be able to kind of ah recreate a past session.

29:18.66
Jon
Right.

29:30.90
Michael
um So that was a big part of what we were trying to do there was was um build up the the data infrastructure side of it, because obviously it’s it’s sort of foundational to everything everything that comes after it.

29:31.32
Jon
Yeah.

29:39.54
Jon
Right.

29:42.66
Jon
Yeah. Yeah. So let’s talk about the the product you created. um You know, again, I learned about it in the dashboards that deliver book, um but you’ve also, I think, called it a ah data product. I think in the book, they’ve also described it as sort of analytical app, but I’m curious about it because it doesn’t strike me.

30:06.46
Jon
Well, I guess it depends on how we define dashboard, but it doesn’t strike me as like a Tableau dashboard, right? Where you go in and you filter and you’re select. Um, how do you think about the product that you created, how it differs from sort of your more standard dashboard? i mean, it’s doing very different things.

30:26.42
Michael
Yeah, it it definitely was. I mean, it it there’s a lot of overlap there, right? So I’m not going to say that it’s um you know completely disjoint from from all dashboards that exist in the world. They’re all Tableau dashboards. but um I didn’t think of it as a dashboard.

30:44.48
Michael
Maybe I thought of certain components of it as dashboards or dashboard-like, but what I was building fundamentally was a ah website, basically.

30:55.37
Michael
it was ah It was a site that lived on our server that um had a backend.

30:58.90
Jon
a

31:01.74
Michael
It had a frontend.

31:02.58
Jon
Right.

31:03.50
Michael
It had data flow into it. it was It was just a much bigger thing than what I think of a dashboard as being. And that’s not to be reductive about dashboards or people that build dashboards at all.

31:15.86
Michael
It’s just, it, there was a, it was a, there was a very broad scope to what we were trying to do. And it was, it was starting from literally nothing, right?

31:21.02
Jon
Yeah.

31:24.03
Michael
We had no, we had no such tooling to do this.

31:24.34
Jon
Yeah.

31:28.06
Michael
Um, that was, that had been built in house. So we were building everything up from scratch.

31:31.38
Jon
Mm-hmm.

31:33.85
Michael
Um, and so that’s why I kind of, you know, I don’t bristle at the term dashboard. People call it a dashboard. That’s fine. um It just was never kind of how I how i thought of it. It was it was a product that that i we were trying to ship and we had we had we had customers of that product and the the customers of that product were the other engineers on the team, right?

31:46.60
Jon
Mm-hmm.

31:57.13
Jon
Right.

31:57.22
Michael
So it it was Yeah, that’s why I use the term that I use. um

32:03.16
Jon
e

32:03.54
Michael
But like I said, I don’t and don’t get offended if anybody calls it a ah dashboard.

32:06.33
Jon
yeah Yeah. But could people like, did you see people go in? Well, I guess i’ll put it this way. Would people come back to you and ask for ah different additions or filters or, or bells and whistles that they kind of wanted to have?

32:24.54
Michael
yeah Yeah, all the time. And that was actually one of the early kind of metrics for success of the tool, product was to me, like I remember writing this down before we built or shipped anything, but I was like, if we are getting feature requests about this, that means that people are using it and that they want more from it.

32:26.04
Jon
Yeah.

32:42.65
Jon
Yeah.

32:47.81
Michael
So the more feature the more feature requests we get, I think the better we’re doing.

32:48.41
Jon
ah Interesting.

32:52.52
Jon
batter.

32:52.92
Michael
um

32:53.05
Jon
Oh, interesting.

32:53.94
Michael
So we, yeah, there was an endless list of of requests that that people would would would submit and definitely didn’t get through all of them.

33:04.75
Michael
But we tried to get through as many as we could and we tried to prioritize the ones that were the most important.

33:05.05
Jon
Right, right.

33:09.27
Jon
Yeah. so um So we’ve talked about the strategists and and and the engineers. um What about the drivers? Did they ever go in and like play around and like either ask you questions specifically, or did you get to watch them like use the dashboard? Or they’re just like, where I’m driving, you just tell me the right strategy.

33:28.12
Michael
um Yeah, it was a little of both. I mean, the drivers have a lot a lot going on and I think it really depends on the driver, frankly.

33:38.74
Jon
Mm-hmm.

33:39.13
Michael
um there are some drivers who are more um definitely more on the analytical side of things and who want to see more of the data. And there are others that are more, I don’t know, fly by the seat of your pants kind of kind of drivers.

33:48.68
Jon
Sure.

33:56.10
Jon
Yeah.

33:56.58
Michael
And so, um yeah, definitely there were drivers that were using the tools that i that that we had built. And in fact, I remember one sort of exciting time when…

34:10.55
Michael
Basically, in in between any practice run when the car would come in, the driver wouldn’t necessarily get out of the car. um The driver will just kind of sit there while. you know we make changes to the car or we look at some data or we wait for the track conditions to be what we want them to be before we head out again the driver doesn’t pop out all the time and so the driver always has uh an ipad basically that gets hooked on their steering wheel um which which shows them a number of different things um and and we have different views for different tools that that they can look at but um

34:27.22
Jon
isn it

34:45.72
Michael
I do remember one time or but the the first time that um the driver was looking at one of my tools in the car. I was like, oh, that’s that’s pretty cool.

34:53.52
Jon
Yeah.

34:54.53
Michael
That’s that’s that’s fun.

34:54.55
Jon
Yeah, that is cool.

34:55.97
Michael
um

34:56.22
Jon
Yeah.

34:57.05
Michael
I think most of the time the drivers would look at the tool was sort of at the behest of the of the engineers when we were back in the truck and um

35:03.76
Jon
Oh, yeah.

35:06.97
Michael
you know the engineers would be looking at something and it it would be more like a hey look at this kind of kind of thing um that was generally how it went but not for every driver there were there were at least one that uh definitely more kind of data minded i thought

35:11.42
Jon
Right.

35:16.86
Jon
Yeah.

35:20.45
Jon
Right. You’ve got the early Tom Cruise, at the beginning of Days of Thunder, and then the Tom Cruise at the end of the movie who understands a little bit about racing.

35:25.40
Michael
right right exactly yeah yes

35:28.01
Jon
Yeah. um That’s very cool. That’s very cool work. um Before I let you go, um what are you what are you up to what are you up to now?

35:40.31
Michael
um Yeah. so now I’m, I’m kind of back in the, in the cybersecurity space, actually. um As, as interesting as the racing life and world is, um it is not entirely conducive to raising a family of kids.

35:56.74
Jon
Right. Yeah. lot travel.

35:58.10
Michael
ah There’s a lot of travel and a lot of commitment in the summer. And um yeah, we’ve got a, we’ve got

36:04.20
Jon
Yeah.

36:05.50
Michael
One at home and one on the way. So um sort of back into cybersecurity. I’m also doing um some work with InfoGrate, if if listeners are familiar with that, data visualization consultancy, trying to kind of up-level their data intelligence capability, basically.

36:16.41
Jon
Oh, cool.

36:28.50
Jon
is it

36:29.08
Michael
So that’s kind of what I’m doing right now.

36:29.56
Jon
And do you do you do you do you do you still have time on the side to do any little ah motorsports data viz just for fun here and there?

36:38.35
Michael
I’ve thought about it. um you know I still have the Twitter account.

36:39.73
Jon
Yeah.

36:40.79
Michael
I still have the ah ah followers that I had then, i’m sure i who I imagine would be interested in um and seeing that come back to life.

36:42.30
Jon
Yeah.

36:47.77
Jon
Yeah.

36:52.02
Michael
I’m less excited about the platform now than I was in the past.

36:54.26
Jon
Right, right, yeah.

36:54.38
Michael
But yeah.

36:56.21
Jon
Well, there’s a yeah whole new F1, whole new set of cars.

36:57.97
Michael
Yeah.

36:59.93
Jon
you know A lot has changed, so or is about to change, yeah.

37:00.34
Michael
Yeah. yeah

37:03.98
Jon
so

37:04.69
Michael
Yeah.

37:04.92
Jon
um

37:04.97
Michael
so I’ve, I’ve thought about it. I don’t know. i might, I might give that a rest for a little bit, but, uh, it was.

37:09.85
Jon
Well, as someone who’s had those two young kids, probably having the extra time is not something that’s going to find its way to you in the next next few months, but ah but maybe someday.

37:15.45
Michael
Right. Right. Right.

37:18.53
Jon
um Okay, last thing.

37:18.84
Michael
Yeah.

37:20.29
Jon
um If people want to learn more, ah either about your work with McLaren or today in the cyber world or or the stuff you’re doing for InfraGrate, what’s the best way to get ahold of you?

37:31.35
Michael
ah Yeah, you know, I don’t have any sort of website or anything that I can i can direct people to. But if you want to collect ah connect with me on on LinkedIn, I’m i’m Michael Gethers on LinkedIn. um

37:42.52
Jon
Awesome.

37:42.97
Michael
Yeah, I’ll be happy to happy to say hey.

37:45.55
Jon
That’s great. Michael, thanks so much for coming on the show. Really interesting stuff and and and good luck with the ah new baby.

37:52.02
Michael
Thank you. I appreciate it. Bye.