00:00

Patrick:

The Mining Your Business podcast is back with yet another solo episode now. How are you doing, Jakub?

00:05

Jakub:

I'm doing very, very nice, Patrick. Thank you.

00:08

Patrick:

I'm not sure about you, but in my free time, I totally daydream about process mining and all the things that it could be applied for. Don't you?

00:16

Jakub:

No?

00:18

Patrick:

From realistic to plain silly use cases, we have five process ideas that are a little bit more unusual than your typical P2P. Shall we get into it?

00:26

Jakub:

Let's do it, man. Hello, everyone. Patrick, so I'm a little bit hung over, and you know what it means.

00:42

Patrick:

It's time for a solo episode.

00:43

Jakub:

It's time for a solo episode, exactly. This is usually if you're listening to us for the first time, we are not usually recording hungover, but sometimes it just happens. And since Patrick, my dear colleague, is leaving to our US office actually in like two days from now, right?

01:02

Patrick:

Yeah, two days.

01:03

Jakub:

Yeah. So I think he was celebrating with, with the German office. Unfortunately, I am from Czech Republic and also live here, so I couldn't attend. But Patrick, how was the farewell and especially how did you enjoy being here in Germany again and now flying back to the US?

01:18

Patrick:

The farewell is tonight. I had a different engagement last night, so it's back to back, which I am looking forward to, but probably not tomorrow. To answer your question. Yeah, it's, it's exciting, obviously. The going back to the US office and seeing all the dear colleagues over there again and spending some quality time with them.

01:39

Jakub:

Once you get there. So forever one who listens to us again for the first time or doesn't know us, we have an office in Austin, Texas, and I've been visiting Patrick before. It's first of all, an amazing city, great people in there as well. And what I wanted to say, Patrick, have some tacos for me. 

01:59

Patrick:

You know I will.

02:00

Jakub:

Yeah, great tacos in Texas. Really nice. If you never had that, you should probably go to Texas just for that. I mean it.

02:08

Patrick:

Exactly you know, I knew that as soon as you didn't have a script prepared for the intro, you're going to start rambling about tacos. I saw it coming.

02:17

Jakub:

You know, everything just ends up at Taco However, I did not write the script that much. I also wanted to say that I'm actually going on holiday in one day. So we are kind of getting ahead with the schedule and with recording before Patrick leaves and everything. And yeah, we wanted to come up with something a little less serious than we usually are. I mean, we were talking about acquisitions of Microsoft recently. We had people from Skoda telling us about their process mining initiative. We talked about Python and all this, all this stuff, and we thought, you know, let's take it a little lighter this time. Let's look at something interesting, something that doesn't really fit the fold that much. And we thought for a very long time actually, I had this idea in my head that we should do an episode about some weird, interesting use cases that you wouldn't really think of because everybody's just so preoccupied by you know, the business processes.

03:24

Patrick:

P2P or order to cash, you know, the usual stuff. Yeah.

03:27

Jakub:

You know, and then I was just talking to Patrick and pitching him this idea. So, dude, do you want to do an episode on accounts payable or do you just want to talk about crazy use cases in process mining? And he just looked at me. Are you even asking me this?

03:46

Patrick:

Yeah, I think that was the fastest decision I've ever come to.

03:50

Jakub:

So if you guys are interested in accounts payable, I think you have to wait a bit. It's in the pipeline. It's coming. But I think this was an easy easy choice, I think. And, you know, so saying that we will dive into five interesting use cases that we think process mining could be applied to. It goes from why don't we do that yet to, ok, this is probably stupid, but it's just so much fun that we want to discuss it anyway.

04:22

Patrick:

Exactly. Why aren't they doing? Ahh, that's why they're not doing it.

04:26

Jakub:

Also the difficulty could be all you know through the roof for some cases but we will talk about it anyway because we just want you to get your imagination also, you know, through the roof. So we picked a couple of criteria’s that we will be discussing for each of these use cases and to tell you a bit more about this. So we will start with the obvious question why is this interesting? Obviously not everything you want to put into process mining, you know, analysis or dashboards. And that's why we want to kind of brief you in on why this use case that we picked could be interesting. And what added value could it actually bring?

05:07

Patrick:

Absolutely. And it should also be noted that the five things that we're talking about came from you guys. So we got some suggestions and yeah, this is based on that. We obviously added a little bit of our own, please get in touch and give us more ideas. We always like to incorporate this into our shows.

05:28

Jakub:

Exactly. So the next criteria will be like, what problem could it solve? Because at the end of the day, you don't do process mining just for the, for the sake of it, right? You want to kind of get some interesting insights into problem that you're trying to solve. The next one, that's probably more technical, what would be the case ID?

05:48

Patrick:

Yeah, so what's the one thing that we're following through the process? What's the one thing that's building a process? What the activities are relating to? What's the central core case that we're looking at?

05:58

Jakub:

Exactly. That would be your invoice, your purchase order, your sales order, or you'll see. Obviously, once you have the process mining in place, you want to take a look at what would it actually bring, what would you measure, how would you build your reports. So that's definitely interesting. And what might be also interesting are the dimensions. Again, you will hear our ideas about that because if you're looking at the standard business processes, you will split it by business organization, by company goals and whatnot. And in our use cases, the splits might be even more entertaining.

06:40

Patrick:

Yes. And of course, there are some challenges with these use cases of course, they're a little bit out there. So one, there's an unfamiliarity with getting data into these structures, right, that we're not familiar with. And obviously there's some overall difficulty associated with just how you get the data. That's always a challenge with the typical ERP system. This is fairly easy, but with the use cases that we're going to be talking about, this could be a little bit more difficult to do. And obviously, the last part, who would pay for it? Because at the end of the day, having a process mining initiative is a big endeavor, right? So obviously it needs to pay off or someone needs to fund it. Somebody needs to, you know, think that this is worth to do. So the question is, who would pay for it?

07:27

Jakub:

Yeah, but we will leave this questions, especially with the overall difficulty and with the challenges to academia. I know that there are a lot of academics listening to us. Maybe they will get some ideas for their thesis and, you know, who would pay for it. We all know that there are free process mining tools also in the market. We interviewed at least one of the founder of such a tool. So yeah, just go learn Python and use process mining tools that are already available.

07:53

Patrick:

Absolutely. Shall we kick it off?

07:56

Jakub:

We shall Patrick. What's the first one?

07:59

Patrick:

The first one is American Football, the kickoff to touchdown or however you like to call it. We were thinking about looking at the American football genre and in there looking at individual ball movements. Can you explain a little bit?

08:15

Jakub:

Oh, yeah. So first of all, this idea actually came from our colleague, Nicolas Möller who was also on the podcast all the way back, all those episodes back. But I think this is for me, this is like I would say, no brainer in terms of why this is not a thing yet. So if you ever followed American sports and be it basketball, be it baseball or American football, you would quickly find out how crazy Americans are about data. They are analyzing everything. Seriously, everything.

08:49

Patrick:

Yeah. Yeah, exactly. This is the third dunk from a person named Marcus who has a grandmother called Dorothy, who was born in Connecticut or something like that. They have stats you can't believe, right? So there's a ton a ton of data. And also why I think it's so interesting is that there's also so much money in this sport, right? So, yeah, and teams are already spending tons on data analysis and trying to get the edge if you can kind of see a quality or an insight that could give you like a 2% edge in some play. I mean, wouldn't you do it?

09:25

Jakub:

100%. And I just wanted to mention, speaking of this data, I remember last year in the last season, there was this I think a guy played a guy with the same name. There was a quarterback and then some defensive player. And there was a statistics that this is the first time ever that the guy with the same names got like all these, you know, fumble and the tackle and everything that the defensive player just did everything for the quarterback, which is just crazy, crazy likelihood. Yeah. Anyhow, speaking on why it's interesting. So I think I was trying to do a little research on that and I didn't find any paper or any study on application of process mining in American football. I did find a study on European soccer, on a European football, but not American one. Why? Well, one of the reasons I was thinking was that maybe process mining kind of rooted and started in Europe and is just getting to US. So maybe they just didn't jump on that just yet or maybe it's just not really out on the internet, who knows. However, the interesting part for me is that, as we mentioned, the guys in the US, the analytics, they love their data. And I think all the data is pretty much already available. And how I picture this is that you would basically track every action in offensive action in in the league. So what would be the case ID? For me it would be the ball movement. You know, if you ever watched American football, a quarterback picks up the ball or doesn't actually that's also one of the cases. Also can happen. And you would basically track whatever happens to the ball from the moment the referee or the snap starts. So from the moment the center snaps the ball to the quarterback all the way until it's end of the play. And you know, this is when it gets interesting because when you're attacking basically have three downs. So this is one of your first dimensions you would see like what is your first down happy path, what is your second down happy path, what is your third down happy path? And then you snap the ball, your process starts and then you see, ok, quarterback picks up the ball and then all these variations start like in 70% of cases he just hand it over to his running back. However, in 30 or 25% of cases he's actually throwing the ball and then the tree just goes crazy because if he's throwing the ball he can actually, you know, he can be tackled and go down. He can also throw the ball and it will result in an incomplete or he throws the ball to hit wide receiver who picks it up and just, you know, runs for the touchdown and then, you know, your ideal path is that every action should end up with a touchdown. Right? That's the perfect scenario unless you have one minute to go in the game and you're up 20 and you just want the time to run out.

12:39

Patrick:

Yeah absolutely. But I think this is also a really, really interesting because in this case I think the outcomes or weighing the outcomes of these specific paths is incredibly important because there you might look at, ok, I have 99% go through this path but it's the 1% outliers where for example the center snaps, quarterback fumbles, the ball the defense recovers and you have the touchdown scored on yourself right or you have a safety or something, something that doesn't usually happen but it is detrimental to any drive in the game right so it's that 80 20 rule but like a little bit more specific. So looking at these outliers and maybe also what contributed to these outliers could be very interesting to look at.

13:23

Jakub:

Yeah. I think if you were analyzing then the dimensions that you're looking at and I already mentioned like that the number of down that you're playing, but also a quarter if you know, American football is played in quarters and as you are closing out on the game, the stakes get higher and you know, the pressure gets on. So maybe it would be interesting also to look at, ok, maybe against this specific team on a home game. They are specifically vulnerable at the end of second quarter because their focus just goes down and they just want to go to the locker rooms already. And maybe if we expose them for this specific action because, you know, their defense is already worn out, this could be resulting in a higher conversion ratio. And this is ultimately this 2% edge, Patrick, that you were mentioning. And these dimensions I can think of so many.

14:20

Patrick:

It's so interesting because you wrote down QB RB wide receiver and I think it's a really good point because there's such a vast amount of combinations of players that you can put in. Right so I mean your quarterback's usually going to be the same person, right? But you can always see when players are a little bit you know, they've run like seven downs already and they're tired. You switch them out. Right? What combination of players and what combination of plays works well against this particular defense setup? Right. So there's a whole bunch of variability in every single play that you do.

14:51

Jakub:

Exactly.

14:52

Patrick:

But thankfully, there are so many plays that are happening every season that you can probably form or synthesize some sort of trend out there and figuring that optimum combination, that optimum play to run in that moment in these weather conditions in your home stadium or away. Getting all these dimensions and figuring out if I do this, I have a 2% better chance of making it seven more yards than if I were to do this play. No brainer.

15:19

Jakub:

It would be super interesting. It might be difficult to translate this into like follow up seasons because the teams just change so much. But you could make a point that if you were looking at, say, at the performance of a quarterback, of a specific quarterback and just measured let's say that you only measure his performance over the years and see how he evolves. You could see like from year to year whether he incline to running a specific play, whether he inclines to, let's say, you know, in process mining, you can measure the throughput times, whether the time that he holds the ball after the snap is after a certain point, it's just detrimental to the result of the whole play. Maybe he just should do something with it a fourth of the second faster or something.

16:09

Patrick:

Yeah, exactly. And that could also lead to maybe saying, well, your wide receivers aren't being getting open down the field. So he has less or more time to really find one that could result in more sacks or he needs to scramble out of the pocket more. And there's a whole bunch of leading or, you know, effects that come from a such a simple calculation, ok, my throughput time from snap to throw is longer in my QB than most teams? That has consequences.

16:36

Jakub:

Oh man. I mean, I would love to see process mining use case on NFL. I honestly think that this is the future.

16:48

Patrick:

Of course you do.

16:50

Jakub:

Please don't quote me on that. If you think that this is also a great idea and eventually put some effort into it and trying to come up with something just let us know because I would so much love to see this and then I would be like, I told you so guys, this was a great idea.

17:10

Patrick:

Now, we should probably talk about the difficulty as well. I mean, it is actually surprising how much data there is out there. I mean, you can track player stats all the way from their college days all the way to right now. The games also have up to date really real time data that is being recorded about these games. And obviously not to mention all the replays that you can watch to really, you know, scrub the data and actually get this stuff out of there. So the access to data and I think it's already in a pretty good spot because like we already mentioned, data analysis does take place for large teams and things like that already. So this wouldn't be that big of a stretch to do, I think.

17:52

Jakub:

Yeah, I 100% agree. Again, the obsession with the data is already in there and we would just feed them with yet another way that they could look at it. And again, if you saw the movie Moneyball, it's actually about baseball. The concepts still apply here. It's again, just a new way of looking at the data and who's to say that this wouldn't be successful. And again, in these 2% edge or maybe even less is still an edge that in such a competitive environment where the teams are basically all equal and that in any point of time, any team can beat any other team. We see it every season. This could make all the difference. So in my opinion, it's just a matter of time before somebody introduces it. And then everybody would be like, why didn't we do it earlier?

18:45

Patrick:

Yeah, absolutely, absolutely. So I think that there's definitely a use case for it. There's definitely a strategic value from it.

18:54

Jakub:

Before we move to the next one, I also think that this idea of process mining could be applied to many sports. Actually, I can think of tennis, you know, every, every serve. I'm sorry, I'm so good on tennis terminology, but every serve could be a specific process in soccer maybe as well. So if you let's say if you start with every game or something and see what do you end up with, you know, could be also interesting.

19:21

Patrick:

Absolutely. I mean, there's some that are not suited like track and field, like the hundred meter dash, like gun goes off and there's a throughput time of 9.86 seconds before you're in the goal and yay, that's it.

19:33

Jakub:

Anyway, I love that the data analytics is just so much rooted into sports and that sports is producing so many interesting datasets, it's crazy. And I'm just looking forward for these technologies to be applied there as well.

19:47

Patrick:

Absolutely. Shall we move on to the next one?

19:50

Jakub:

Let's do it. I think we spent a big chunk of our episode talking about NFL.

19:54

Patrick:

Yeah. So the next topic we wanted to talk about, and that's also dear to my heart, is music production, right? So what does music production really entail? Well, it usually involves an artist. It involves some sort of a recording session and some sort of mixing and studio, you know, magic and then some sort of release, but usually coupled with some marketing, some merchandise and things like that, you know, that's I don't know, I'm being very, very reductive here, but that's usually the core of it. Right. So why is it interesting? Well, we would like to know. I mean, every big artist or every artist in every studio would like to know, hey, I'm spending a lot of money making the song. I would like it to be successful. Is there anything that could tell me if a song will become successful or not?

20:45

Patrick:

That's why it's interesting.

20:46

Jakub:

So what you're saying is that you would be trying to optimize your way of releasing your music so that it makes the biggest possible splash on the market.

20:58

Patrick:

Correct. So, I mean, this is obviously a very subjective field. Music is very subjective in its nature. So figuring out the release strategy is one part, but also song structure, right? You could have that as dimensions like what type of song am I writing? Am I writing a big love ballad? Am I writing a summer jam? Am I writing a country song? You know, it could be anything, right? But in figuring out what to do in these specific cases, to make it a success, is pretty big.

21:33

Jakub:

So it should basically confirm my theory that if you release a Christmas song in the middle of the summer, you might not really succeed.

21:41

Patrick:

Hahah, yeah, I think you're on to something here. I think we need a little bit deeper on that.

21:50

Jakub:

Dude, it's just july right now.

21:52

Patrick:

Exactly. So get ready for Jakub's Christmas jingles.

21:56

Jakub:

Yeah, I think maybe we can find a loop in the market. Maybe if we just do that, people get excited and they're like, I like that. So it's just six months since Christmas, actually. Well, seven but.

22:07

Patrick:

Well, I mean, then again, this is one of those things, you know, the cliche, you go into a supermarket in November or October, and they're already starting to blast Christmas music. Right? It's not just, ay, it's December. It's time for Christmas. No, no, this starts way beforehand, right? And if you want to crunch out a turn out a Christmas song, when do you need to start by latest? Like, how long does it take to write a Christmas song? I mean, this is one of those things you really want the artist to take the time with it or you just want to crank up the most formulaic Christmas jingle you've ever heard that old trend, right?

22:41

Jakub:

Yeah. It would be also interesting to categorize this into different buckets. First of all, I mean, the music style, that's the obvious one. But also like, you know, in standard business processes, you always have these amounts. So you know, what is the total amount of an invoice or something. In here, you could, for instance, measure a number of listeners or another dimension that comes into my head is like, whether the music group is already established or whether it's their first ever song. What market are they in? I think that the song production will be very different, you know, from Germany, Czech Republic, all the way to, I don't know, Vietnam or something. So these types of inputs would be vital for the production. You could also, for example, look at what production company is helping you with because maybe some are more likely to succeed than not. Obviously, that would probably go hand in hand that the bigger the music group, the more established the music group, the more the listens they get because they already are on the market. But it will be very interesting to see are those rare cases, those rare unicorns that produce the first song and the song just goes viral. This is exactly what I would like to analyze. Like, what is the magic? What are the magic beans for this? Right. What is the formula? That some songs just got it and some didn't. And you could argue that the difference between them is nonexistent. On the other hand, you might be missing everything else because, you know, music production is a complex topic. It's not just the song, it's everything else. If you have a strong brand, strong social media, you know, you take a picture with some famous guy on Instagram and your song is suddenly going to be trending or, you know, we are living in an in our times of reels and Tik Toks and whatnot. And you never know if some famous influencer is going to put a stupid video with your song in it and your song is just going to go crazy.

24:49

Patrick:

Yeah, absolutely. And also, I think you'll be surprised to find how many artists have had a failed music career under a completely different name and just paid a lot of money to rebrand and redo and then start over and then make it big. So there's there are some pitfalls here, and I think a lot of it has to do with them. Of course, money like how much money can you stick into this? You know, there's a difference between somebody from a more established in the music industry family, and that has the connections versus someone that's making music in their bedroom right now. These are two completely different starting points.

25:26

Jakub:

So Patrick, who would pay for it?

25:31

Patrick:

Well, I mean, if you could, if you told the production company and say, I have this tool that would let you analyze your entire catalog and with a 90% chance you'll probably hit the billboard 100 if you find the formula, then you know, who wouldn't want to pay for that? I mean, there's tons of people that would love to crank out nothing but formulated songs that kind of reach the Billboard Top 100. I mean, there's a lot of money in that.

25:54

Jakub:

Yeah, give us million dollars and will do it for you, and the data, and some data scientists as well.

26:00

Patrick:

I mean, the only challenge is, of course, breaking down such a subjective thing into data. Because why do you love Careless Whisper, Jakub?

26:13

Jakub:

Why do I?

26:14

Patrick:

Exactly. Who doesn't love Careless Whisper? You can't really put your finger on it, right? It's not really quantifiable in a way. There's some markers, of course, what chord progressions and things like that you used, but in the end, you don't want to make a tool that just says, build this one song in this exact same formula. Because songs are also about exploring and making new things so formulaic songs are also maybe not the best thing.

26:42

Jakub:

Yeah. So this is music production. I think we should move to the next point. Before we do, Patrick, I'm just going to tell you that I'm not I'm never going to dance again.

26:51

Patrick:

I'm so sorry to hear that.

26:54

Jakub:

All right, up next, farm yield. And this is actually the something that we have been discussing in our second episode about what is process mining to certain degree, right? So if you remember our what is process mining episode which interestingly is our most viewed, most listened to episode ever by far like it's crazy.

27:17

Patrick:

Our magnum opus, you could say.

27:19

Jakub:

Yeah, yeah, you could say that. Yeah. That's a very nice word. Thank you, Patrick, for being so smart.

27:25

Patrick:

Oh stop it. Ok, so farm yields. Why talk about farm yields? Why is that interesting? And what do we mean by farm yields? Right. I just kind of slapped that on there as a title. But what I'm talking about is the whole farming industry. We have a limited amount of land. We have limited resources, we have limited water. We want to reduce pesticides and all these things. But at the end of the day, we want to get as much crop yield from our farmlands as possible. So food production is incredibly vital to any functioning economy in my opinion. That's not really a controversial thing. So having that down and figuring out what results in the best yields is obviously a very, very interesting point.

28:11

Jakub:

What would you actually track in this case? Would it be like the seed?

28:17

Patrick:

I mean, the seed would be interesting just because you could see very interesting things when you do it on a seed level, the problem is just data collection at that point. So what I was thinking you could track plots of farmland, right? Because farmers have different plots of land that are on different hills and, you know, separated by fences and things like that. And they grow different crops on different fields. Right. So you could track a plot of farmland. And there you could track the sun exposure, how much you water it, the rainfall over the season, the pesticides that you use, the timeline of it all. So when you plant the seed, because I don't know if you know this, but I mean, farmers have like so much knowledge about how they get the most out of their land. But it's like such old timey knowledge a lot of the times like things that were passed down from generations to generations. So when you see three crows crossing from east to west at 5 a.m. on a Sunday in Easter or something, that's when, you know, you need to go out and plant some seeds. So this is like really, really old knowledge that is probably accurate, right? They know what they doing. They've been doing it for hundreds of years. But maybe there's more quantifiable data that tells us when the optimum time is to really plant the seed.

29:33

Jakub:

Patrick, I do come from a long line of farmers. So I know what you're talking about. I did work in agriculture when I was young.

29:42

Patrick:

Yeah, you did?

29:43

Jakub:

Yeah, yeah, yeah. It was good times, good times. But I completely get it and what I love is the time dimension. So how long is the optimal way under what circumstances to get the highest yield. Should you already like best decide today or should you just wait another day or two so that you actually achieve this optimal performance. And I'm sure that there are like stations or research laboratories that are looking at this in a very lab like conditions but with process mining, if it was possible to track this with some what some data set and this actually brings me to the challenge because it's not so easy to track the plots of farmland and, you know, compile it in a digestible data set. But if you could do that if you figure out some smart way on how to how to tackle this, you could have a sample in the real life and not really in the laboratory. So you could actually start making these points about the sunshine, about the rainfall and so on. And look, ok, so maybe if I on this farm one, if I switch every two years, the type of crops that I'm seeding here maybe is going to result in something better than what I've been doing now with changing it every year. You could also start looking in some, you know, data like what type of crop would you change there? And, you know, eventually just you're looking to optimize the way of what you're going to get out of it. And again, I think this itself makes that the case for our farm yield process.

31:38

Patrick:

Absolutely. Absolutely. It's one of those things that I think some people do and I've seen them do this or talk about it at least where they say, yeah, this grape was grown on this hill and it was harvested this and then but the grape that's on the next hill tastes like so much different because it gets more sun in the evenings and things like that, right? If you believe it, if there is such a difference and it really does make a difference in the yields and the flavor and all these things, then it does really make sense to really go that granular into your crop setups.

32:11

Jakub:

Yeah, I still bet that at the end of the day you would have like this old guy who's been farming for 40 years or something like this is, say this is bull crap.

32:20

Patrick:

Yeah, this is complete garbage. Let's throw it away. I'm just going to watch the birds to tell me when to go out. And sow my seeds.

32:26

Jakub:

What is this? Are you using technology to, to create super, you know, Bordeaux wine or something?

32:33

Patrick:

Exactly, it goes to show like you don't want to obviously replace that knowledge, right? That's worth gold, right? That's what's been feeding us for the last hundreds, thousands of years. So you would likely want to complement that skill sets with tools like these.

32:42

Jakub:

And again, farming has been around for most as long as humanity or modern society, which actually allowed it to be.  I'm not going into history because then I could very quickly could be exposed on my lack of knowledge.

33:07

Patrick:

Yeah, let's not go there. Let's not go there.

33:12

Jakub:

The point is that farming has been a focus of optimization of humanity for so long. You know, you had all kinds of improvements over these hundreds of hundreds of years and the technology hit it also pretty hit pretty hard. I mean, it's not that long when over 50% of all the older population was working in agriculture. And now, if I'm not mistaken, in developed countries, it's like less than 4%, probably even less so, which is crazy because, you know, through the technology advances, suddenly we don't need to send hundreds and hundreds of people onto the farmland because all we need is one optimized machine or some tractor or you know, you are we are starting using drones and why not taking it to the next step with the data analytics? I'm sure that there is already a bunch of ways that they are analyzing their crops and maybe process mining could be next in line.

34:21

Patrick:

Are that be such a cool idea to have an automated drone schedule that just goes and takes pictures of your crops and analyzes like how much they've grown and things like that?

34:29

Jakub:

Dude, this is already a thing.

34:31

Patrick:

Let's do it.

34:32

Jakub:

Out of interest I know that there are companies that like get these data from the satellites and they can like analyze some scenarios of countries and how well the crops are going to go that year because of how the visuals look like. And then they can make like some bets on it in terms of I don't know trading or something, you know, what's going to be a price of a bushel of corn or so.

35:04

Patrick:

Oh wow.

35:05

Jakub:

And this is just crazy, like there are so many possibilities with the data and what you could do it that that the just I think my tiny little brain just can't really, you know, work with it.

35:15

Patrick:

I think it's also important thing to mention, you know, due to climate change, the crop yields and things will be more of a thing. You know, the soil will be stressed harder. The crops will need to go through higher temperatures. And so, you know, having that extra edge that might get you through some incredibly dry periods or something would also probably be really, really helpful.

35:34

Jakub:

But at the same time, look at the countries like Netherlands. I think Netherlands is one of the biggest producer of vegetables in the world.

35:46

Patrick:

Netherlands? Really? I learn so much on this podcast.

35:48

Jakub:

Dude, farming. You know this is because they are using the technical advancement for their own good and maybe they're also using process mining who knows what's Wil van der Aalst is up to in the Netherlands.

36:07

Patrick:

I wouldn't be surprised. We're watching you! Shall we move on? All right so this one's a bit of a somber one. We were thinking about in the field of sociology maybe looking at the life to death process would be fairly interesting. I mean humans are such weird creatures and they go through a whole host of different activities. You could call it. Right so different parts of their life, you know, school and work and you know, you marry or you have a partner or you don't. And you know, that results in different paths in your life. There's a whole I mean, it's incredibly complex. Right? So there's not one thing that you could really point to and say this will predict your life or anything, but there are probably a lot of common denominators in all of our lives that we might want to look at.

36:58

Jakub:

What would the happy path look like?

37:01

Patrick:

Oh, well, that's again very subjective. You probably want a very long cycle time, you know, as long as you can get it. Yeah, that would be a good happy path.

37:14

Jakub:

Yeah. This is this is just really an idea, a pretty wild idea that it would be pretty cool to observe some relations. Like what does happen when you do a certain turn in life. And I'm not necessarily talking about, you know, taking or doing some hard drugs, but maybe, how likely you are to end up, I don't know, studying in a university based on not necessarily the country because there are obviously statistics for that, but in a city or in, let's say, how many brothers or sisters you got to have to do a certain step in life to, again, maybe do university, the success ratio and stuff like that. If you don't look at it from the white perspective that you're really tracking everything that happens from the moment you're born until the moment you die and you look for some specific sociological use cases of analyzing the life key choices and the impact on them on or impact on your overall happiness or likelihood of becoming something specific or actually not ending up in some specific situation. This could be very interesting.

38:31

Patrick:

Absolutely. Absolutely. I think one of the things that would be interesting would be looking at rework rate, like how many times do you go through specific activities? And I know we were joking about this before the show, but if you reach the death activity and you accidentally do that twice because they mistakenly buried you or something, I think that would be a funny outlier to see.

38:55

Jakub:

Yeah, 100%. We have here a Mr. Jesus who died and then he just started to walk again. What's up with that? Can we look at that case? At the age of 33, huh, it's interesting.

39:07

Patrick:

Crucified and then three days later emerges from a cave. I mean, that's crazy.

39:11

Jakub:

So I guess when you crucify a person, they just gonna....

39:14

Patrick:

OK, let's stop where this is going, haha.

39:19

Jakub:

We told you this is going to be a wild one.

39:21

Patrick:

Yeah, exactly. I think it's a very interesting use case. Just I mean specifically, if you would, if you're looking at it from a governmental side, if you were to imagine how do I get my constituents to study. A well-educated populace is beneficial to everybody when they study medicine or anything else, really. So how do I get them to do that? Is there a specific sequence of when you introduce kids to math or something that kind of determines what fields they're interested in? You could analyze this forwards and backwards a thousand times.

39:58

Jakub:

So imagine you starts using this to optimize taxes or something.

40:02

Patrick:

Oooffff.

40:04

Jakub:

By the way, speaking of taxes and government, one of the use cases that didn't make the cut was actually my idea on looking at the at the polls. Whenever you're choosing the new polls, and I was joking. I guess now it's not so funny because of the war and everything, but you know, when you have polls in a country with a dictatorship, then you could just look at those manual touches after the polls are closed. So here are a box of the votes was opened, manually touched and there were some changes from A to B, what's up with that?

40:57

Patrick:

Yeah I'm not sure dictators would appreciate that level of transparency.

41:04

Jakub:

Yeah, but it still cracked me up.

41:07

Patrick:

Shall we go on to our final use case?I know you're waiting for this one.

41:12

Jakub:

I am waiting for it.

41:13

Patrick:

You were texting me last night and giggling.

41:18

Jakub:

So I'll be honest, I really went wild on this one. Don't be offended. Maybe we should make this episode explicit. I don't know. Let's see about that. It's a dating process. So we all are subjected to dating other, you know, men, women, whoever. And I think it would be so interesting to just analyze this process from start to end. Have the, like, use the dataset of all the dates that are happening throughout the world and try to find and just look in the patterns and it would be just so, so funny to look at these things because, you know, your case ID would probably be the date or a partner. It depends really. I think the date would be, you could make points for both because the partners and you see how many dates he went on and so on, but if it's just a date.

42:11

Patrick:

I think that be interesting because well you could look at how you go from if you date someone, how that goes, right? But what if you date multiple people at the same time?

42:24

Jakub:

But then you would have like multiple cases for your person. So imagine that your dimension, there would be, you know, Patrick Bogner and you would select the cases you know, Patrick Bogner goes on dates, and then the case ID would consist of you and the case ID of your partner and you would have like multiple cases there and you could see how successful you are there. You would see that one and only case that you have, right?

42:49

Patrick:

Right, haha, I knew this is going to end up in ridicule.

42:58

Jakub:

Hope you're following us on this one. And I had so many ideas on how you could measure this. I actually came up with this idea yesterday in a cafe when I was after work preparing this episode and there was a table of two young ladies sitting next to me and they were just giggling the whole time. They were drinking, I think, some drinks and everything. And at some point they just went on Tinder, the dating app, and they were just scrolling through guys and making stupid comments and I just couldn't stop laughing. And then I thought, this is a great idea. You could make a swipe to sex process out of this.

43:33

Patrick:

Swipe to sex process, haha. As funny as it is, I think it brings up a good point. You know, dating and and marriage and just general partners in life. I think we could all agree that most people seek companionship in whatever form that may be and having some sort of analytical insight about what determines the success and failure of these partnerships is, I think, from a sociological point of view, very interesting, and maybe not call it swipe to sex, haha. 

44:14

Jakub:

Patrick, I think you're trying to trim this too much, haha.

44:19

Patrick:

Yeah, I'm trying not to have the explicit tag on Spotify!

44:21

Jakub:

But again, the throughput times, what is the impact of someone waiting with a certain activity for certain time? The dimensions, you could finally make a point whether money or looks actually matter or whether it's something else.

44:40

Patrick:

I can take this one - they do.

44:44

Jakub:

How would you know?

44:46

Patrick:

Ahh, you know, being so wealthy and good looking, I can definitely confirm it's true. I can't believe you put birth sign on this list.

44:55

Jakub:

It was the first thing that came to my mind because I was just discussing it with someone the day before and she was like asking me what my birth sign is? And I thought, does it even matter?

45:07

Patrick:

But that would be so interesting though, imagine that it was actually significant, right? Yeah. Maybe Taurus really goes well together with Libras. Who knows.

45:17

Jakub:

Or other things, obviously, age is the easy one. The common interests you have together. So maybe, maybe if you are just like the other, maybe it's not the best thing to date or, you know, who knows? Or the source of the date. I already mentioned Tinder. Maybe you would find out that dating colleagues would not be the best thing.

45:37

Patrick:

I think it's also interesting for governments, right? Because you can see it in a lot of governments having population declines. Right? And there are initiatives to try and get people to date and you know, have kids and things like that. So there's an actual use case for this in in a way. How do you get people to date? How do you get people to successful dates? As funny as this is, this is actually, I think, a very good use case.

46:04

Jakub:

Yeah. I think by talking about applying dating in process mining, is not going to increase our chances on this, you know.

46:14

Patrick:

So I think one of the challenges that I wanted to highlight because this this came to my mind as I was reading this and as you were texting me, I was thinking about tracking this data. I mean, this is obviously a massive invasion of privacy. I mean, also, if you just do a personal tracking, just imagine that and sorry, honey, sorry, I have to interrupt this session, I must log my kiss in the BI tool. It doesn't really set the mood. 

46:43

Jakub:

Yeah, you will just anonymize it and it would be fine.

46:47

Patrick:

Yeah, of course, of course. People would love their government to have some sort of information about their very intimate details.

46:55

Jakub:

Oh, man. I really, really like this use case, and I think it would just be hilarious. However, we also came up with some other use cases that didn't really make the cut into discussing them into the depth, such as some of them are actually already being implemented. There are the environmental processes like what impact you supply chain has on a certain environmental criteria. We already mentioned the government wasting our money. What's the process of that.

47:29

Patrick:

There's a lot of interest in the processes of government wasting our money.

47:35

Jakub:

Especially coming from the governments. Yeah, I mean, I would love to actually talk to someone who's applying process mining in a public sector because it's, you know, it's hugely inefficient and I think it would be great because the data is also there. I mean they're using ERP systems too and analyzing this thing would be tremendous. And it's very, very insightful compared to other private companies that are doing everything possible to optimize their processes. Maybe governmental subsidiaries could also think about something like that.

48:12

Patrick:

Yeah, for sure. I think anybody who has been in and I don't want to dig at anybody here, I'm sure there's reasons for this, but in the German governmental system or signing up for a when you go and live somewhere new or something, you have to change your address. And that whole process you wouldn't think should take as long as it does, but it does. Right. And it's very frustrating to do, right. You have to waste 4 hours of your day waiting in line, getting a ticket and all these things. And you're wondering, why is this so inefficient? Right. I'm sure there's some reason, but I'm fairly certain, since there is no economic incentive really to optimize that, a lot of optimization has been left in the dust.

48:51

Jakub:

Yeah. And I mean, we've all been there. You know, you are waiting in a queue at some public space and then, you know, finally it's your turn and they say, yeah, we don't do this here. You have to go to the other building on the other side of the town where you're going to, you know, take this form and fill it out. And then you can come back to us because before that we can't really help you. Like, are you kidding? I’ve just been here for hours waiting in a queue and you telling me I have to do it over and over again. It's just ridiculous.

49:20

Patrick:

Absolutely. Yeah, it's one of those things that also, I think drives a lot of people crazy. Just the amount of paperwork, you know, we're living in 2022 and we're still doing a lot of things by paper. When we really shouldn't. Well, one, it's wasteful and it's also inefficient.

46:36

Jakub:

That goes for some OCR tool.

49:38

Patrick:

Oh, well, you, we know the trouble with OCR tools.

49:43

Jakub:

Anyhow, what I'm trying to say here is that there is just so many different things that you could be measuring and applying process mining for. I recently saw this, how is it pronounced? Pull, pool, poll?

49:57

Patrick:

Poll.

49:59

Jakub:

Sorry for that. A poll on Celonis LinkedIn profile where they were asking people how long they've been in process mining for and there was this very small subset and less than 1% of people or, you know, respondents who said more than ten years, then about 5% was between five and ten years. And then everyone else was in process mining for less than five years. I'm starting my year five actually in a week. And what I'm trying to say here is that there will be more and more and more use cases where process mining is going to find its way to because it's just a unique way of looking at the data. The companies already understand that, and if we think about it, almost anything that we're doing in our daily life is some kind of a process. How you're doing your shopping, right? How you cooking?

50:53

Patrick:

I was I was thinking about this yesterday, actually, when preparing for this episode, I was thinking, what if we do like a cooking process? But then I realized it's very close to like a production process in a company. You need to check that you have stuff in your warehouse a.k.a your fridge. Then you need to go to the purchasing department, you know, go to the shopping and actually buy the stuff that you need. And yeah, it turned out to be very close to what companies are already doing.

51:17

Jakub:

Exactly. And almost any electronic device. I know that some of the companies are already experimenting with implementing these process mining capabilities into electronic devices as a starting point so that they can, you know, if something goes wrong, they can actually look into the logs and find out what is happening there. And I think it will be more and more present in any dataset, in any way that we work with the data. And it will just be this this category that will accompany us on any data analytics journey that eventually process mining will be implemented wherever and whenever it makes at least a little sense.

52:00

Patrick:

Yeah, you nailed it. You nailed it. And I think this is where you dear listeners come in. We would love to hear your ideas. Right. So we have five ones that we, that we just came up with, but we would love to hear from you. What are your interesting takes on what process mining could be used on? Because everyone has some hilarious idea or maybe some really, really interesting one that just hasn't been done. So we would love to hear from you and love to see what you can come up with.

52:26

Jakub:

Yeah, definitely. Tag us on LinkedIn and write us what is your idea? We would love to hear that or just comment under the under the link that we are posting ourselves or just write us an email, miningyourbusinesspodcast@gmail.com And we would very, very much love to hear from you. However, this is the end of the episode. I hope you had at least half as much fun as we did.

52:52

Patrick:

Doubtful.

52:54

Jakub:

This was brilliant. And if you made it all the way through up till here, thank you for that. And yeah, thank you for listening. Leave us a review on any tool that you're using and we'll be looking forward to hear from you. And we'll be here back in two weeks time with yet another episode of Mining Your Business episode, thank you. Thank you so much for listening and bye bye.