Looking for Starter Kits? Head to the Community Gallery! Now formatted as YXIs for easy installation.

Data Science Mixer

Tune in for data science and cocktails.
Alteryx Alumni (Retired)

Specializing in the scientific study of frightening leisure activities, Mathias Clasen, director of the Recreational Fear Lab at Aarhus University, dials in from Denmark to share how fear can be quantified and data used creatively to understand our ability to adapt to real-world danger. 








Cocktail Conversation


There are definitely unique fears that only your fellow data people will understand. What's the most frightening thing you've had happen with a dataset or a project? Did the resulting fear turn out to provide, as Mathias says, "an adaptive learning moment" for you?


Join the conversation by commenting below!


Mathias cocktail conversation.png





Episode Transcription

SUSAN: 00:00

It's that time again, the spooky season is upon us. [music] A recent survey found that 86% of Americans are as or more excited for Halloween this year than in previous years. Of course, that survey was commissioned by a major chocolate company, so I'm not sure if that's a trick or a treat. Are you one of those people who gets excited at all things spooky? Or do you find those people kind of mystifying? This is Data Science Mixer, the podcast featuring top experts in lively and informative conversations that will change the way you do data science. I'm Susan Currie Sivek, Senior Data Science Journalist for the Alteryx community. We have a special treat for you today for this most frightening time of year. And whether you love or avoid scary stuff, I know you'll find today's guest fascinating. Let's hear all about Mathias Clasen.

MATHIAS: 00:56

An associate professor in literature and media at the English department in Aarhus University. And that's a big research university in Denmark. I'm also the director of something called the Recreational Fear Lab, which is a research unit dedicated to the scientific study of frightening leisure activities.

SUSAN: 01:15

Also, we're hoping to get this episode out right around Halloween here in the US, so that will be great timing, and I think will be a lot of fun for folks to hear about. So, yeah, would you mind also just sharing with us which pronouns you use?

MATHIAS: 01:28


SUSAN: 01:27

Awesome. Mathias told us all about his research on recreational fear. Whether you enjoy or steer clear of scary experiences yourself, I know you'll be fascinated by his lab's methods of quantifying fear, and finding and combining data on this topic. Plus, we'll talk about how you can use fear to your own advantage in your life and career. We might need a dreadfully good drink for this one. Brace yourself with whatever beverage you need for a wild ride through the data-focused study of fear. As you're about to hear, I did not come prepared with a tasty drink. And as you probably know on data science mixer, we often enjoy a special drink or a snack, time appropriate as we're working during the day. So 22, you have anything special there with you right now?

MATHIAS: 02:18

I do, and I was sort of hoping you wouldn't notice because I'm having a beer, and I know it's early where you are. But it's late where I am so I think it's okay.

SUSAN: 02:25

I think it's great. I mean, this is truly-- it's always 5 o'clock somewhere, right? So we're good to go.

MATHIAS: 02:31

All right. That's what I thought. Yeah, so cheers.

SUSAN: 02:33

Yeah, cheers. Excellent. I am. Yes, being boring here at 11.00 AM and having some icy cold water. So it's really not as fun as your selection.

MATHIAS: 02:40


SUSAN: 02:43

It' settled. Good deal. All right. So yeah, maybe you can just start us off with a little bit of discussion of your research generally, especially the Recreational Fear Lab. I love the name, I love the idea of studying fear quantitatively, I think that's really fascinating. So, yeah, I'd just love to hear about how you got into that area of research, what kinds of things you're studying, and maybe just generally speaking, what some of your methods are?

MATHIAS: 03:06

Right. So for a long time, I've been studying horror, mainly horror literature, and horror movies. And I think through a combination of luck and stubbornness, I was able to turn my personal passion and enthusiasm for scary movies, and scary books into a relatively lucrative professional pursuit. And then a couple of years ago, I got lucky again and got to meet the right collaborators with whom I managed to get funding for the Recreational Fear Lab. So it's actually funded by the Danish funding body that funds basic research.

SUSAN: 03:44


MATHIAS: 03:44

So for now, it's a three-year grant, and I have some really good people employed in the lab. And the purpose of the lab is to really get a quantitative grip on this phenomenon that we call recreational fear. And the way I see recreational fear, I think it's more or less a term that we coined. And the way I understand it is as a kind of broad sort of diffused concept that includes horror but is broader than horror. So one of the things we're doing at the moment, and this is something I'm really excited about, is investigating the prevalence of recreational fear in Danish daycare institutions.

SUSAN: 04:24

Oh, wow.

MATHIAS: 04:24

And that probably sounds a little bit crazy, because how would you find frightening activities in nurseries and kindergartens? But we thought this is something we want to check out. We had a hunch or maybe an intuition that quite possibly Danish kids are being frightened for fun. And so, we had two research interns in the spring in the lab. And they did a wonderful job of collecting pilot data and found that it's really, really widespread. Those poor kids are being frightened witless before noon anyway. Most of those recreational activities that we have discovered take place in the morning before they have their nap.

SUSAN: 05:06

So what kind of-- what forms are those taking?

MATHIAS: 05:08

You'd think not only that I was crazy, but that the whole Danish population was crazy, if I told you in detail.

SUSAN: 05:14


MATHIAS: 05:14

It ranges from, normal or probably more or less universal pursuits such as hide and go seek, and chase play, switching off the lights and reading mildly scary stories. But also, local nursery rhymes, which are really terrifying.

SUSAN: 05:33

Oh wow.

MATHIAS: 05:34

They center on cannibalism, and monsters eating kids, and they often end with a jump scare. And also, more kind of imaginative forms of pretend play. So taking the kids into the woods, and letting them know that they're hunting for a troll or whatever. So that was a fun surprise to see how widespread those phenomena are. And also, to learn that the teachers who instigate those recreational fear activities in nurseries and kindergartens see it as a sort of-- they see it as their duty to expose the kids to manageable and controllable doses of fear in a fun sort of context.

SUSAN: 06:13

Huh. That's so interesting. And so how are you actually studying or measuring the fear that is being caused or the outcomes of it?

MATHIAS: 06:22

For this study, we haven't started doing actual empirical research yet. So far, those results are based on interviews with the teachers. And that's because of the pandemic, we couldn't really-- we wanted to do some observational studies to begin with, but I'm hoping this will turn into a more ambitious project that would, I imagine, probably rely on behavioral observational data. So we would look at the kids, and look at face expressions, and body posture and so on when they're being frightened by the adults, and also maybe self-report. I mean, we have found self-report to be a very reliable measure of fear in other studies, but I don't know how good kids are at giving an accurate indication of their own fear levels, but that's something we'd have to pioneer.

SUSAN: 07:10

Yeah, yeah. And that's again, what I think is so interesting about your work is figuring out ways to quantify these experiences and emotions. So I want to come back to actually the issue of the pandemic here, just a little bit. But I feel like I would be remiss as an interviewer if I didn't go back to you talking about your personal fandom as far as horror goes. So what are some of your favorite horror movies or books?

MATHIAS: 07:32

Huh. That's the sort of question I should be able to answer without thinking because I get it a lot, but it's actually surprisingly difficult to answer. And I used to not like horror. When I was a kid, I would steer clear of it. And then the sort of reversal happened in my, I think, teenage years, and that's fairly common that that sort of developmental window opens at that point.

SUSAN: 07:56

That's interesting.

MATHIAS: 07:56

But Halloween would be one of my favorite movies. A classic John Carpenter slasher from 1978 [crosstalk]. And anything by Stephen King, basically.

SUSAN: 08:07

Sure. Yes. The classics. Excellent. Very cool. Awesome. So getting back to these questions around measurement and so forth, you talked a little bit about self-reports, and gathering data on people's experiences of fear. So I noticed, looking at your body of research so far that, you do use quite a bit of survey data in your research. And I'm just curious what some of the challenges have been in that process? How you've been able to glean insights from those survey data. Anything that might have come up for you as you've worked on gathering that information?

MATHIAS: 08:37

I think one of the coolest studies we did, and this is what I did with my-- he was my PhD student at that point, Ian [inaudible] [Jensen?], and we teamed up with an American personality, psychologist John Johnson. And we wanted to get some basic data on how common it is to be a horror fan because nobody really knew and so, we did. And that was a survey in which we used crowdsourcing; I think we used Mturk. And so, we paid respondents to answer a bunch of questions having to do with their relationship with horror. So, do you like horror? Do you like horror books or video games? Do you go to haunted houses? How often do you watch horror movies? Do they make you more or less anxious? Do you watch them with others? And also, a whole battery of personality questions to get their personality profile. Because we wanted to see if there was a unique horror fan personality?

MATHIAS: 09:32

So that was one study that generated a lot of data. But happily, John is a real wizard. And so, he did the statistical analysis. But I think one challenge in that approach was that we could only recruit people over the age of 18. We did get, I think, a representative sample only, adults though. And so that left a gap. But that's one of the gaps we're trying to fill now, to look at a more broad developmental spectrum, and to figure out when the appetite for being scared for fun emerges, how it changes reliably across childhood and adolescence. But other than that, I think using a survey for that study was the right way to go, and it gave us some really fascinating insights.

SUSAN: 10:20

What was your favorite finding from that study?

MATHIAS: 10:23

The finding that made me the happiest, and that made the headlines in those stories that were written about it was that horror fans tend to be-- they score fairly high on a personality dimension called openness to experience.

SUSAN: 10:41

Oh, yeah.

MATHIAS: 10:41

And I was happy to learn that because horror fans face a lot of prejudice. They're very often seen as people of poor taste, and kind of immature aesthetic preferences. But here we actually found that people who enjoy horror tend to enjoy intellectual stimulation. So that was a fun finding. But, probably okay, that's probably my second favorite finding. The one favorite finding is that more than half of the American population based on this study actually enjoys horror. So it's not a niche phenomenon. 55% answered yes to the quote, to the statement, "I tend to enjoy horror media."

SUSAN: 11:19

Interesting. Yeah. And I like how you created suspense there, too, actually, by counting down your second favorite and your best favorite.

MATHIAS: 11:27

Right? That was inadvertent but yeah.

SUSAN: 11:30

Nicely done.

MATHIAS: 11:30

Thank you.

SUSAN: 11:31

No. That's very cool. That's interesting to hear. My husband is a big horror fan. I'm more of the sci-fi end of things myself, but he really is, cinematically savvy, and a literary guy and enjoys these things from his own perspective, which I totally appreciate, so.

MATHIAS: 11:48

Right, exactly. Yeah. And horror can be a lot more intelligent than many people give it credit for.

SUSAN: 11:54

For sure.

MATHIAS: 11:54

So there are horror movies that are morally interesting, aesthetically interesting, psychologically probing all of that.

SUSAN: 12:01

Yeah, yeah, absolutely. So along these lines of gathering data in new and interesting ways, another one of your studies, I think the title was, Playing With Fear, you actually had a number of different types of data. There were physiological data. There were the survey data, and then also video footage. I think that's such a neat and creative combination of data to use to study, in this case, the haunted house experience. So can you tell us a little bit about how you went about designing that selection of data sources, and how you actually combined all of them to arrive at some insights?

MATHIAS: 12:35

Yeah. That was a wild study. And I think it almost killed our postdoc.

SUSAN: 12:41

Oh, wow [laughter].

MATHIAS: 12:43

But it ended well, and we got the study published in Psychological Science, which is just about the coolest psychological journal.

SUSAN: 12:51

Yeah. Congratulations.

MATHIAS: 12:52

Yeah, thank you. We're really happy with that study. But actually, so it came out in, I think, in November 2020. But the data was collected in 2017. But it was such an amazing wealth of data to put a positive spin on it, that it took a long time. So what happened was that we had established a collaboration with a commercial haunted attraction here in Denmark. And in 2016, we did the first study in which we went to collect data on fear regulation strategies. We wanted to learn how the guests regulate, overregulate, or downregulate their own fear in their pursuit of the optimal experience. And the next year, we wanted to do something even more ambitious because that first study from 2016 was based only on survey questionnaire data. But for the next study, we mounted surveillance cameras at three different spots in the haunt. And we still have a huge plastic bag full of those network cables because they had to be connected by cable.

SUSAN: 13:58

Oh my gosh.

MATHIAS: 13:59

So we have, I think, about a mile of our network cable in one big tangle just waiting for the right intern haunt circle.

SUSAN: 14:07

Oh, no.

MATHIAS: 14:09

Yeah. So we mounted the cameras, and we mounted infrared lights so that we could film in the dark because we want that behavioral data. We wanted to be able to look at the guests when they got a jump scare. We designed questionnaires in which we asked about their experience. How much fun was it? How frightened were you? Do you remember what happened in this room? How frightened were you in that room? And then we used our heart rate data. So these little lightweight heart rate monitors that we would mount on people's chest before they went in and then take them off when they came out and download the data onto a computer. And then we had to synchronize all of that data to look at the exact point at which they entered the haunt. And it was just a nightmare, but a nightmare with a happy ending because the results were really fascinating.

MATHIAS: 14:52

And for that study, I think the result that that created again, that created the headlines was that we found in both self-report and heart rate data, we found a sweet spot between fear and enjoyment, an inverted U-shape. I think you can imagine if you plot fear on an x-axis and enjoyment on a y-axis. If people don't feel a lot of fear inside a haunted attraction, they're bored and they don't find it enjoyable. If they feel a lot of fear, they get overwhelmed and they also don't enjoy it. But just the right, the sweet spot of fear, that's what people aim for. Which is in a way surprising because so many horror movies and horror video games and so on, they're marketed on their ability to scare people, the scariest horror movie ever. But actually, people don't want maximum fear. People want just the right amount of fear. So that was one fun thing to come out of that study.

SUSAN: 15:44

Yeah, yeah, absolutely. So maybe a little bit of a tangent. But I'm curious on what your thoughts are on that, the super extreme haunted house here in the US, whose name I'm blanking on, the one that people have to sign all kinds of waivers to go through.

MATHIAS: 15:57

Yeah, that's McKamey Manor in California.

SUSAN: 15:59

Yeah. Is the attraction there, do you think primarily just that people are curious to see if it really can fulfill the promise of extreme fright, or--

MATHIAS: 16:10

Yeah, I think so.

SUSAN: 16:10

--there is something else going on there.

MATHIAS: 16:12

Something else. And I think once you look at those extreme haunts, do you enter a domain that is maybe partially overlapping with, but also largely separate from the ordinary recreational horror experience? Because I think it becomes about something else. And I'm not sure what that is because I haven't studied it, but I think the people who are attracted to those extreme experiences are not the average horror fan. That would be my guess, but it's something we would like to look at some point.

SUSAN: 16:44

Yeah, yeah. It just came to mind as we were discussing this.

MATHIAS: 16:48

Yeah. No, I wouldn't do it personally. I wouldn't. No way.

SUSAN: 16:51

No. No, thank you. I'll pass.

MATHIAS: 16:53

Yeah, exactly. Hard pass.

SUSAN: 16:55

My goodness. Yeah. So I think this is all really interesting. I love the idea of combining all those different data sources, despite the challenges to really get these deeper insights. One, other--

MATHIAS: 17:06

Yeah. Also, because--

SUSAN: 17:06

I'm sorry. Go ahead.

MATHIAS: 17:08

No, no. My apologies for interrupting. But the video footage provided its own challenge in that we had to hire two independent coders to go through all those hundreds of hours of film and code facial expressions for joy, and surprise, and fear, and so on. So that in itself was challenging, but also really unique. And that's the sole reason we keep going back to that haunted house. Partly because if you're interested in recreational fear or just ordinary fear, you have to convince the ethics committee that [laughter] what you doing is scientifically sound. And they will not give you the thumbs up if you plan on inviting participants into a lab and giving them a jump scare, and measuring heart activity. But you can go to a haunted house where people have chosen to visit, and where they have a really intense experience more intense than what you would see at a horror movie.

SUSAN: 18:02

Yeah, yeah. Interesting. Yes. The good old IRB, the good old ethics review there. It could be challenging there. Interesting. So yeah, the other study of yours, sort of the real-world horror, I guess, that was super interesting was this recent study that you co-authored on how horror fans and people who are inclined toward those more kinds of fear inducing experiences, that they may actually have been more resilient during this pandemic time. So I'm curious how that research came about, and what some of your major takeaways were there.

MATHIAS: 18:35

Right. Yeah, that was a very fun study to do, and one that really made the rounds around the globe. Actually, a career peak for me was when Heather Langenkamp, the actress who was in the original A Nightmare on Elm Street movies, tweeted about the study.

SUSAN: 18:51

Oh wow.

MATHIAS: 18:51

That was really cool.

SUSAN: 18:53

Very cool.

MATHIAS: 18:54

But yeah, but the origin of the study itself is sort of interesting because some years ago, I think at least five years ago, maybe more, I was interviewed by a journalist for a science magazine called New Scientist, on scary stories and their roots and cautionary tales and fairy tales, and so on. And I'd said something about my idea that maybe horror stories can be adaptive in a biological sense. That it makes sense for us to be attracted to frightening, imaginative simulations because maybe we get better at handling real-world danger. And so, in the early days of the pandemic, this is back in March or maybe April in 2020. This journalist reached out on Twitter, and I think sort of tongue in cheek asked me if I thought horror fans were doing a better job of dealing with this, what seemed sort of apocalyptic, the whole pandemic and the lockdown? And I thought to myself, I'll bite. Sounds like something that could be fun to investigate. So I teamed up with some good people, including Coltan Scrivner, who is an American scientist who studies what he calls morbid curiosity. And my old PhD student and John Johnson again. I think this time we used Prolific, which is another platform for recruiting participants.

MATHIAS: 20:11

So we recruited a couple of hundred of participants, and asked them a bunch of questions about again, their relationship with horror. How many horror movies do they watch? How many apocalyptic movies, alien invasion movies, zombie apocalypse movies? And also, we designed, or we adapted a survey instrument for getting a measure of pandemic resilience if we wanted to see how well psychologically our respondents were doing during the lockdown. And so, we found actually what we were hoping to find, a correlation between the number of horror movies people watch and their psychological resilience. So people who watch a lot of horror movies were actually doing a better job of avoiding some of the negative effects that many people felt, such as problems concentrating or sleep problems. And so, it's of course, it's a correlational study, we can't prove any kind of causality. But I think it makes sense that people who watch a lot of horror movies have a lot of practice in the fine art of emotion regulation. I mean, if you watch a really scary horror movie, you will have to somehow regulate your own fear. And maybe those techniques that you use, whether it's self-destruction or reminding yourself about it's not real, it's not that dangerous. It's not real. Maybe some of those strategies can be used in the real world. And so, in that way, I think the findings made sense.

SUSAN: 21:39

Yeah, yeah. That's so interesting. And to your point about experiences of fear being adaptive, it reminds me of the Danish nursery kids being taken out into the woods and told about trolls.

MATHIAS: 21:50

Yeah, yeah. And I think that's what the teachers intuitively assume that it will be good for the kids to give them these little doses of fear. There is an emerging scientific focus on the usefulness of risky play and adventurous play for kids. Because parents over the last couple of decades have done a wonderful but also horrible job of protecting their kids. So that they don't get a chance to play on their own, they don't get a chance to climb trees and fall down, and really identify and challenge their own limits. So that sort of helicopter parenting may actually be doing the kids a disservice. Because they don't know what fear feels like. They don't know what it feels like in their belly when they climb too high. They don't know how to regulate their own anxiety.

SUSAN: 22:39

Right, right. That's interesting. Anything else out of the pandemic study that you thought was particularly significant?

MATHIAS: 22:47

Yeah, there was one other thing that I thought was really interesting, and that was our finding that people who watch a lot of what we call prepper movies. So movies about the end of the world, zombie apocalypse movies, and alien invasion movies, they felt better prepared for the pandemic. And I wasn't that surprised. I mean, I think it's about 10 years ago now that the Center for Disease Control in the US, they had a zombie awareness campaign, which it started as a joke, an April Fool's joke. But they created this campaign for fun with the tagline, "If you're prepared for a zombie apocalypse, you're prepared for any disaster." Yeah, I think that's one of the few times in the history of the CDC that people really paid attention. So people went to their website to check out this campaign. And they and they sort of meant it because the way in which you would prepare for a zombie apocalypse is not that different from the way in which you would prepare for a hurricane. Maybe, except for the amount of guns needed to tackle either situation. So my thinking is that if you watched a lot of those prepper movies, you will have simulated scenarios of social collapse or great social upheaval that are not that different from what you could witness in the real world.

SUSAN: 24:04

Yeah, absolutely. It's so interesting. So I'm curious if there are things that you've learned about studying people's experience of fear, that might inform data scientists who are doing work with data in business kinds of settings? Maybe the ways that you're studying emotion or things about consumer experiences, has anything come to mind there?

MATHIAS: 24:25

Well, immediately the thought that comes to mind is the appeal of fear. I think that, at least to some people it's counterintuitive that we would enjoy fearful experiences. But on the other hand, I think most people recognize the lure of recreational fear. And so that might be something that could be applicable in domains of marketing or advertising. And of course, also maybe some of the ways in which we measure fear and enjoyment could be applied in a more kind of business domain. I think one of the things we've found in those haunted house studies that we keep doing, I mean, we're doing one, we're designing one right now. And we did a really fun study last year on the different motives of different horror fans. This year, we're looking at interoceptive ability and fear, so people's ability to register the signals that come from their own body, and the degree to which that is correlated with a fear response. But we have found consistently that self-report is pretty reliable. I mean, if we ask people, "How afraid were you on a scale from 1 to 10?" And they give us a number, and then we'd look at their physiology, I look at their heart rate, what they say actually matches what their heart tells us. So if somebody in the business world is interested in getting a measure of fear, it may not be necessary to go out and buy heart rate monitors, you can just ask.

SUSAN: 25:54

It's interesting because I think we often kind of pooh-pooh self-report as unreliable but maybe this is one area where we did have a reasonably good sense of reliability.

MATHIAS: 26:05

I think so. I mean, of course, you have to take into account that sometimes-- one thing we found in the personality and horror study was that there were very small gender differences, much smaller than people have traditionally assumed, so men and women enjoy horror. Men and women want horror to be scary. The one gender difference, the significant gender difference we found was in the degree to which people are willing to admit that they were easily frightened by horror movies.

SUSAN: 26:34

Oh wow.

MATHIAS: 26:34

So, you ask a man, "Were you frightened by the movie?" He's more likely to say, "nah, of course not," than a woman might be. Biases, of course, are worth taking into consideration. But the way we've dealt with it is to design these. We've printed out questionnaires and make sure that people fill them out in private with nobody else looking so that they don't feel the demand to live up to anybody else's expectations.

SUSAN: 26:58

Right. Right. That social desirability bias. Interesting.

MATHIAS: 27:00


SUSAN: 27:01

Cool. So I am curious, have you learned anything about fear from your research that has had an impact on your life in any way?

MATHIAS: 27:09

One thing I have learned is the variety of coping strategies that are available. I mean, I've gotten better at regulating my own fear. And so, one study we did, we found that people use three different kinds of coping strategies when they try to manipulate their own fear. And they are cognitive, and behavioral, and social. And by far the most predominant ones are cognitive. So that's people manipulating their own thinking, and that seems to be also the most effective ones. So if people are watching a horror movie, and it's really scary, just think about something else or remind yourself it's fiction that seems to be fairly-- it's more effective than covering your eyes, which is a behavioral strategy. And also, I think this emerging picture of recreational fear as something that can be beneficial. Something that can help you build resilience, something that can function as a kind of stress inoculation, and something that may even have bonding effects. I mean, people very often seek out frightening media, haunted houses, horror movies in groups, and it looks like they might actually feel increased bonds with the other group members. That's impressed me, these positive effects.

SUSAN: 28:27

Yeah, interesting. Yeah, I can imagine people who have-- maybe data scientists who are giving presentations, thinking about fear in a different way as a result of this conversation. That's cool.

MATHIAS: 28:38

Right. Yeah, we all feel fear just prior to giving a presentation. But I think mild doses of fear have the effect of keeping us focused and making sure that we don't lose focus on what is important right now. At least that's what I tell my students when they have to give a presentation, and I can see their heart racing and so on.

SUSAN: 28:56

I know, we've all been there.

MATHIAS: 28:58

We have.

SUSAN: 28:59

So I believe you have a new book that has just been published and could you tell us a little bit about that?

MATHIAS: 29:05

Right? It's called A Very Nervous Person's Guide To Horror Movies.

SUSAN: 29:09

Love it.

MATHIAS: 29:10

And it came on Oxford University Press. And it's supposed to be an easily read roundup of all the best research on the psychology of horror. And it's aimed at people who are interested in horror movies, but also sort of nervous about them. And so, I tackle these different kinds of nervousness that people might have in relation to horror movies, for instance, can they kill me? Can they be bad for my mental health? Can they be immoral? Are they bad for kids? And then I look at what the science says and offer some science-based advice on how to survive a horror movie.

SUSAN: 29:47

Awesome. That sounds great. And I think that's also perfect for the time that we'll be putting this podcast out, so folks need to check that out and get ready for scary movie time around Halloween also.

MATHIAS: 29:57

For sure.

SUSAN: 29:58

So one question we always ask on the podcast we call this the alternative hypothesis recurring little segment here. The question is, what's something that people think is true about working with data, using data, and the kind of research that you do, that you in your experience has found to be false?

MATHIAS: 30:18

The immediate thought that comes to mind, and this is probably my bias because my background and training are in the humanities, where many people have a kind of instinctive knee-jerk reaction against quantification. And so, I come from a tradition in which quantitative research is viewed as reductionist, something that really strips away the richness and the beauty of whatever phenomenon you're trying to investigate. I think that's wrong. I mean, I think the only way to really get at a mechanism, to really get at the, I guess, the causal underpinnings of a phenomenon, even an aesthetic phenomenon like horror movies, is through reduction and quantification. And you can study such a phenomenon using data, using surveys, and video recordings, and heart rate data and so on, without stripping away any of its richness or indeed paradoxical beauty.

SUSAN: 31:19

Yeah. As an English major and medieval and renaissance studies minor from undergrad, 100% agree. Yeah, I've been using all of those angles, to me as well just enriches the understanding it. I think that's fantastic. So very cool. So one thing that I didn't ask you earlier, but I was curious about it as you were describing it, you were telling me about the coding of the video footage from the haunted house and how there was a lot of manual coding going through all of that. And then, of course, trying to reach the [inaudible] agreement and all that good stuff. Did you ever consider or would you maybe in future use a machine learning type of method for that where it would be automatically coded for you?

MATHIAS: 31:59

Yes. And we did make contact with-- I met a guy at a party--

SUSAN: 32:04

[laughter] It's the best yeah.

MATHIAS: 32:06

--who does that sort of thing, and who was really interested in getting that data because he thought he had some algorithms that could do that sort of thing for us. But at that point, the coders had already done their job, and they had done a good job. But yes, that is something we want to do in future studies, for sure.

SUSAN: 32:23

Cool. Yeah. Yeah, it's a really interesting area that has a lot of potential consequences, using machine learning to study emotion. It's something I've been kind of fascinated by.

MATHIAS: 32:32

It's something, actually, we're working on something new, but it's kind of secret. So we'll have to do a follow-up a year from now if we get the funding. But that involves machine learning, and psycho physiological measures to design a virtual reality horror experience that shapes itself after the user's physiology. So blink rate and galvanic skin response and facial expression, so that once people start getting bored the software will throw in a jump scare, and vice versa. Once they become almost overwhelmed, it will maybe turn off the lights in the simulation, to keep people in that sweet spot of fear that we found in the haunted house.

SUSAN: 33:12

Oh, it's so interesting. Can we use that brief description that you just provided, or is that secret too?

MATHIAS: 33:16

I think that's okay.

SUSAN: 33:17

Okay? It's sort of to be sure. Yeah, that sounds awesome. So yeah, I know that's been an interest in media, of course, for some time, having some sort of real-time adaptive experience for people, so that sounds very exciting.

MATHIAS: 33:29


SUSAN: 33:30

Yeah, if we'll get that on the calendar for a year from now, I love it.

MATHIAS: 33:33


SUSAN: 33:37

[music] Thanks for listening to our data science mixer chat with Mathias Clasen. Join us on the Alteryx community for this week's cocktail conversation to share your thoughts. Let's talk about your scariest data science moment this week. You've got a compassionate audience here. There are definitely unique fears that only your fellow data people will understand. What's the most frightening thing you've had happen with a data set or a project? Did the resulting fear turn out to provide as Mathias says, an adaptive learning moment, for you? Share your thoughts and ideas by leaving a comment directly on the episode page at community.alteryx.com/podcast, or post on social media with the hashtag data science mixer and tag Alteryx. Cheers.




This episode of Data Science Mixer was produced by Susan Currie Sivek (@SusanCS) and Maddie Johannsen (@MaddieJ).
Special thanks to Ian Stonehouse for the theme music track, and @TaraM  for our album artwork.