Data Science Hangout | Jennifer Listman, Statespace | Culture that Helps Avoid Burnout

Hi, everybody. Welcome to the Data Science Hangout. If you're joining for the first time, it's great to meet you. I'm Rachel, I'm the host of the Data Science Hangout. If this is your first Hangout, this is an open space for the whole data science community to connect and chat about data science leadership, questions you're facing and really what's going on in the world of data science. I'll also say, as I mentioned last time, if you are watching this on YouTube in the future, you can also join us live and we'd love to have you. But we always want to create spaces where everybody can participate and we can hear from everyone.

There's three ways to ask questions. So you could jump in live, you can raise your hand on Zoom, you can put questions in the Zoom chat and just put a little star next to your question in the Zoom chat if you want me to read it out loud. But we also have a Slido link where you can ask questions anonymously too. And someone from RStudio will help share that in the chat in just a moment. But I just want to reiterate, we love to hear from everyone, no matter your level of experience or area of work.

So today, I'm so excited to be joined by my co-host, Jenny Listman. Jenny is a Director of Research at State Space. And Jenny, I'd love to have you introduce yourself. I know you've been on here a few weeks before, but share a bit about your work.

Hi. Well, so first, thanks so much for having me. And thanks to RStudio in general for providing this forum for the community. It's great. So I'm Director of Research at State Space Labs, and I think probably most people don't know of the company. Some people may know of some of our products, but without knowing the name of the company. So it was founded by neuroscientists who met at NYU. It was founded about five years ago. So the company combines neuroscience, gaming, and AI to build products that help people maximize their performance and connect with other people while having fun. So that's vague. But the underlying technology is based on research methods from sensory motor neuroscience and cognitive psychology.

So there are multiple products that the company has. If you're a gamer, in particular, if you're an FPS gamer, you might know of our first product, which is called Aim Lab, which is a first-person shooter assessment and training scenario, although a lot of people use it just because they think it's fun. We also have Aim Blocks, which is Aim Lab in Roblox. We also have two esports and gaming coaching platforms where we can connect players with coaches and training content. And we also have a digital health group that's been working for a while in stealth mode. So I think most people have no idea, especially people who are familiar with the company through Aim Lab or Aim Blocks, they have no idea that we have a digital health arm. So the digital health group is working on gamified cognitive and motor performance assessments.

And so the company, while it's founded, it was founded by scientists and the underlying methods and the underlying analytical methods and assessment methods or strategies are developed in-house by scientists. Many of the other people who work at the company come from the gaming industry or are diehard gamers. So the engineers, the designers, user experience, marketing, and analytics, there are people who either came from the gaming industry or are really serious gamers. We also have former esports pros and coaches, streamers working at the company.

Business problems and data at Statespace

So the company has, I would divide our data questions, problems, goals, into kind of two buckets. So first of all, I should preface that by saying a lot of people assume that because Aim Lab is free, I think the default assumption is that we're selling people's data, which we're not. We just keep all of the data in-house and use it to improve the product or build other products. So that's number one. But I would say we have two different buckets of data and analysis needs. So one is the kind that you might encounter at a company, another company with a customer-facing product. So retention, customer segmentation, marketing analytics, A-B testing. So we have those needs. But there's also a lot of science going on at the company. So we have another bucket of data needs that might be more what you would expect at a digital health company or even in an academic lab.

So that even applies to the Aim Lab product or Roblox product, so even the gaming products and not just the digital health products. We're still doing things like figuring out how do we measure certain aspects of human cognitive motor performance through their interactions with a computer? How do we extract scientifically valid signals from the data? How do we estimate the extent to which different environmental factors or practice habits affect performance? So, yeah, when I think about our data needs, I sort of bucket them into these two categories. So we have a data and analytics group that works on the problems that you might normally associate with a data science or analytics group. And then we have an R&D group or a science team that works on a lot of the other problems, but the results of our work feedback into each other.

Resume tips and self-promotion

Something I'd love to ask you about from a follow-up previous session was around resume reviews. And I know we had brought that up about how it'd be great to be able to connect people from the community, people that could share their resume and others that could help provide feedback. And I'm curious if there is specific feedback or tips that you generally share with people?

I think a theme that I've seen on this, the data science hangout, but just in general online among the data science community is like, well, how do I transition? How do I get my first job in data science? And I think part of the problem is just the word data scientist. It means so many different things that at a certain point, it loses meaning, right? I think people use that term to describe a business analyst or someone in another context, someone could just use the same term to describe someone who's using neural nets to solve problems. Right now, I don't do either of those things.

So I think partly the data science community might need to start doing a better job at describing these different kinds of jobs, because if you're just starting out or you're trying to, whether you're transitioning out of academia or you're an undergrad or graduating with a master's degree and you want to go into data science, what does that even mean? So I think it might be good if the community could come up with some better definitions.

As far as general advice, I would say one of the things that I've seen are people who are not, this is what we talked about previously, people who are not good at promoting themselves. And I, in a previous life, was definitely guilty of that. And so I think just that skill alone, learning how to write about your own accomplishments as if it was someone else writing for you, is a skill to learn on its own.

Just that skill alone, learning how to write about your own accomplishments as if it was someone else writing for you, is a skill to learn on its own.

I will mention too that this is just an experiment, but I was thinking if we wanted to start a resume review club through the Data Science Hangout, I did make a Google form where you could just say, you could upload your resume or you could raise your hand to say, I would love to help review someone else's. So I just put that link in the chat if people want to check it out.

Yeah, I would say this is totally my fault. I did this to myself. This came up because, I don't know how long ago it was, maybe three, four weeks ago, someone at State Space forwarded me a resume for someone who was applying for one of our data analyst positions and asking, is this person qualified? Should I tell them to apply?

So I looked at the resume and immediately certain things jumped out at me like the most important accomplishment within each job, that previous job that had been posted, the most important accomplishment at that job was listed last in the bullet points. For more than one job that was listed, this person was, before giving themselves credit for anything, was referring to their colleagues, which is wonderful. I trained under this person who helped me or I contributed to something, something, something, but then they had done things on their own and they weren't promoting that stuff first. And then out of curiosity, I looked at this person's LinkedIn page and saw, well, there's a whole bunch of amazing things that were listed on their LinkedIn page that hadn't even made it onto their resume. So I responded and said, they may very well be qualified to apply for this job, but before that happens, forget about this job. This person just needs to really redo their resume and do a better job at telling everyone how wonderful they are.

Aim Lab tasks and the science behind them

Hi, Jenny. I'm an M Lab fan, FPS junkie. So when I see that there's an M Lab company, I'm so excited that you'll be talking about. So you mentioned that M Lab was started by neuroscientists. I'm curious how the task, basically, for everyone here, M Labs, basically, you have different tasks that you, basically, sensory tasks that you click on. And how was the task first designed? Was it driven by the neurological science side or the gaming side? And also, could you give us a flavor of what sorts of analysis you can do on M Lab data? Basically, I'm trying to imagine, could gamers actually help healthcare research?

Yeah. So they can. So the first tasks and the underlying technology was developed based on the kinds of experiments that often go on in psych labs, except with a much more boring stimulus. Like there's a horizontal line on the screen, and there's a vertical line on the screen, and click when you see this. So some of those basic types of experiments were, basically, overlaid with gaming scenarios.

So, well, I see someone in the chat, not just executive function tasks, but there are, I don't know, maybe at least like 100 tasks right now in M Lab. Some of them concentrate more on tracking something that's moving, like a smooth tracking movement. Some of them are more ballistic movements. Some of them assess working memory. But the trick with something like M Lab is that, like I said, a typical experiment in a lab, you know, people are paid. They're there for, you know, maybe an hour, and it's boring, but they're being paid. So if it's a game, it has to be fun, right? So the tasks in M Lab have to balance the, you know, from the perspective of the scientist, being able to collect scientifically useful data, while being fun enough that people want to play it, and also being similar enough to the gaming scenarios that they're actually trying to train for. So it's a total balance. It's a continuous balance when, you know, for any of the tasks that are designed.

So we have had, you know, research projects that were done directly with, you know, small numbers of M Lab users. But for the most part, people are just playing the game, however they want to play the game, with whatever equipment they want to use. Typically, you know, they're just playing the typically in their home, it could be two in the morning, it could be one in the afternoon, you know, somebody might play it for 10 minutes, somebody might play it every day.

We also, like I said, we now have this digital health group. So we are using tasks, we're sort of taking the tasks from M Lab, and in some cases, porting them to tablet, in some cases, leaving them as a, you know, PC based tasks. We were working with several different universities on projects, having to do with assessing concussion, assessing cognitive changes under hypoxia, so low levels of oxygen, oxygen deprivation that you might experience at high altitude. We have another research project with pediatric cerebral palsy patients. So the same technology that's and gamify tasks that are that were built for M Lab can be used for for these research purposes and to, you know, assess cognitive and motor performance outside of gaming.

What's most exciting right now

So for me, at the current projects that we're working on, I think, you know, Statespace is a young company. So we're at the point now where I, we are better able to take advantage of all the data that that we're collecting. So, you know, we have this enormous volume of performance data, of human performance data. We've been able to write a couple of journal articles based on that. So if you're, if you're coming from academia, the size of our data sets, is really exciting. So, so that, you know, sort of like a playground, if you're, if you're coming out of academia to have access to such a large amount of data. So on a regular basis, I find that exciting. But we're at a point now where we are going to be able to, you know, combine data from the different products that we have, so that they can feed off of each other, to really maximize the experience of our users, and get better at personalizing the user's experience to help them improve. So I think that, you know, it's been like a long time coming. And, you know, before you can do something like that, you actually have to just spend a lot of time immersing yourself in the data, to even figure out what's there. So yeah, so that, that's on the horizon.

Consulting and transitioning to full-time

You had previously had your own consulting company as well. And I know there's a lot of people in the data science space, who maybe even during COVID, just recently started consulting businesses. And I'd love to hear a bit about your experience in that. And if you have some tips that you might share with people listening.

Yeah, so I, my initial academic training was in anthropological genomics. I worked in a psychiatric genetics lab for a number of years. And then when I left academia, the role that I had was in a, a center for STEM education at NYU's engineering school. And I did some, you know, project outcome measures, I instituted a lot of, a lot of quantification of the programs that they were offering to, you know, they, they offered programs to like thousands of local kids in middle school and high school, but weren't necessarily keeping track of the effects that the programs had. So I was able to, to institute that, write some papers on that. I wrote a lot of grants in that role.

And a specific problem came up that I thought I can solve, I can probably solve this with R. And it had been a couple of years since I had had to use R. I think I started using R in grad school, probably like 2004. So there was no RStudio, it was all command line. But I thought I, I, there's this problem and I'm pretty sure I can, you know, find a package and solve it in R. And when I started doing that, I was like, you know, I miss this. I miss this part. Like I, I loved my job. I loved the people I worked with. But I realized like I, I, I need to, I need to do this again. So I left that job and set aside a chunk of time to turn myself into a data science consultant. And to do that, I, and I, I know this is advice given, you know, maybe controversial advice given to people like, oh, assign yourself projects and make a blog. And so, which I did. And I, I do understand that in order to be able to do that, you have to have enough free time to, to be able to accomplish that, which I did.

And, you know, basically assign myself problems that would force me to have to learn new packages, new tools, new analytical methods. And after doing that, I, I landed a, a gig as a consultant for an education nonprofit that had a gigantic longitudinal data set. And I worked for them for about maybe like a year and a half before starting at Statespace. And even for Statespace, I, at first I was a consultant writing a really large grant, like a monster grant. Didn't get funded, but that's okay because it laid, it laid groundwork for a lot of other stuff. And then, and then I was, I was hired full-time at Statespace.

Culture, flexibility, and avoiding burnout

So Statespace is unusual. It's, the culture is unusual. The, you know, the flexibility of the work hours is unusual. The company has been remote from the beginning. There was no real office when the company was really small. And then people started getting hired who were in different time zones. There were thoughts of getting an office, but then the pandemic hit. And in the meantime, the company raised more money and expanded. There are people who work at, for Statespace, you know, all over the world. You know, we manage, you know, in a single meeting, there could be people on like seven different time zones. So, and the, the culture at Statespace is people just have to get their work done. So nobody's really a nine to five employee at Statespace.

And I, I understand that, that, that also some people have negative opinions about a company that's run that way, where, you know, the work hours are like, just get your work done. It's whatever the hours are that, that you need to do it. And, you know, there's no specific vacation time, just take vacation when you need it. And I understand that there are circumstances in which a company might take advantage of that, or where people might work for a company that has that kind of policy where they would feel really pressured to just never take vacation. And, and it's just, that's just not how it works at Statespace.

So I think I've had, there's one job that I had when I left academia. I think I was there for maybe a year. Not, not the, not the job at NYU. It was a different job where it was really nine to five. And I remember the first time I had to, I had a dentist appointment, I had just a dental cleaning, and I had to take a half of a, of a sick day to get my teeth cleaned. And I was just like, I don't, I don't think I can deal with this. And, and I, I totally understand this is just a personal preference. There are plenty of people who want a job where they know at, you know, 5 p.m. or 6 p.m. I'm done. Nobody can bother me. And when it's the weekend, it's my weekend and no one can bother me. For me and my family, this, the level of flexibility that I have at Statespace is actually really important and, you know, a huge benefit for me.

How do you ensure employees at Statespace have a work-life balance to avoid burnout? That's a good question. I mean, it is definitely a risk. So managers have to be really careful and be on top of that to make sure that that doesn't happen. I think it depends on the role that somebody has at the company and the level of, of responsibility that they have. For some of the people in, for example, engineering, or we have a really, really large community team, meaning the people who are on Discord or Twitter communicating, you know, in constant communication with users responding in real time when people have a question or have a problem. And so for those people, you know, certainly there's work to do on the weekends, or if something crashes, there have to be engineers who are willing to, to fix something on the weekends. But yeah, managers just have to be really careful about who has to take responsibility for that kind of thing, like in an emergency.

Yeah. And again, I think that that is an issue that people have with companies that have this sort of, you know, nebulous working hours and lack of specific vacation time. All I can say is, I think it's working pretty well at State Space. You know, personally, I haven't had a problem with it, you know, up to now. I mean, I appreciate being able to, you know, in the middle of the week, if I have to take two hours during business hours to make phone calls to manage my 87-year-old mother's health care, I can do that. I can, if it's, you know, a family member's birthday, I can stop working at three and make a cake from scratch and do whatever I want to do to help them celebrate their birthday. And then I don't feel, for me personally, then I get work done on the weekend because it's really quiet and I don't have any meetings. I don't have any work related meetings. And it doesn't bother me that I'm doing that because I also know that, you know, I have the freedom in the middle of the week to exercise or, you know, do these other things I need to do. But it's definitely a personal preference. There are people who really need specific work hours for their lifestyle.

I mean, I appreciate being able to, you know, in the middle of the week, if I have to take two hours during business hours to make phone calls to manage my 87-year-old mother's health care, I can do that. I can, if it's, you know, a family member's birthday, I can stop working at three and make a cake from scratch and do whatever I want to do to help them celebrate their birthday.

So HR does keep tabs on that. And they will make people take time off. Like, I mean, they can see who's not taking time off. So yeah, occasionally people are just told, like, go disappear for a while. You need to do that.

I know we had to do that at our studio, because we do have unlimited time off. But sometimes when you have unlimited, you don't actually take it off. So we have like quarterly check-ins now, where you have to say, like, are you on track to take this minimum amount?

I like Mark's tips in the chat, too. He said, a good manager who recognizes when the team needs a break is key. Yeah, or a manager who takes time off themselves to set an example. Definitely.

Age, cognitive performance, and sleep

Yeah, I see this question about, you've read an article saying that cognitive abilities begin to decline at 26. And we rely on previous experience more at that age. So is this true? And does it mean I have to try harder to beat my 12-year-old brother?

We did collect some data on age and performance in AIM Lab. And our data did indicate that performance seems to peak before 26, like 24, 25 years old. But I would qualify that by saying that, you know, the data that we're collecting, our data set is messy. This is not under controlled conditions, right? Our users are playing AIM Lab at home. All of their equipment is different. And we have no idea what time of day they're playing, right? We don't collect information on their gender. So it's not like this was done in a lab, but our data showed that there was a decline after like mid-20s. And there have been published studies that were conducted under, with data conducted under controlled conditions that have similar results, like at around that age.

Yeah. But yeah, you know, things like game sense and experience, you know, cognitive control in terms of restraining yourself so you don't shoot a teammate instead of an enemy, things like that. You know, age might give you an advantage.

Yeah, I mean, generally, depending on the task, you do see some like declines from that. It's the pinnacle. But just so that in case anyone's worried, they're probably not going to be noticeable. You might be psyching yourself out a bit too much with it, that we have cognitive test performance data from like eight years to like 90. And for a lot of these tasks, there's not a big difference. Things that involve reaction time, you will see a decline. But things that rely on sort of accuracy or some sorts of, or depending on the memory paradigm, you won't have like a significant decrease, unless there is some neurodegenerative disorder or some sort of injury. But if you're generally a healthy person, you can probably still beat your 12-year-old brother.

Yeah. I mean, the other thing that's important is sleep. So yeah, you need to, you know, I think there are people who think that if they practice, you know, just like cramming for a test, well, if I practice for hours and hours, I'll get better. But if you do that until it's like three in the morning, then you're actually, you know, not, you won't retain what you're learning, right? If you're practicing and training, you're not going to really retain it. And also, you know, lack of sleep in general harms cognitive performance. And gaming is a cognitive task. So yeah, you should get, everyone should get some sleep.

Digital health partnerships and adapted tasks

So right now, we have several research projects that are that are ongoing. One with Mount Sinai Hospital in New York. That's the project that we're doing where pediatric patients with cerebral palsy are using some aim lab tasks that have been adapted for, you know, some are on tablets, some are on PC, but adapted for use in that context. And for that, it's not only that the mechanics of the tasks had to be changed, but all of the imagery and sounds and text had to be changed because there can't be any gun imagery or, you know, anything about shooting like this is a tool that's being used in a hospital and with kids. So the design team had to totally overhaul the tasks and, you know, turn, like remove anything having to do with a gun or words having to do with kill or shooting.

One of the tasks, it's a tracking task. And this is like one of my favorite things that came out of our work over the past year. It's a tracking task. So, you know, someone has to follow something that's moving on the screen and track it. And so instead of targets like round shaped targets, the targets are cats. And they, I don't know, for people who are familiar with video games, the target's health or an opponent's health is, you know, sometimes shown on the screen as like a decrease, a bar that's decreasing in size or like, you know, something that's colored that's fading. So the cat's health is like a rainbow that's basically coming out of the cat's butt. And it just like shortens as the cat's life shortens. So yeah, so we had to adapt the tasks for use for pediatric patients. And definitely our design team had a lot of fun with that one.

And then we have partnerships at Indiana University where we have a large study going on with a lot of the athletes at Indiana University assessing concussion using adapted in-lab tasks. And that's also at Indiana University is where we're doing a study on hypoxia.

Yeah. And we've also partnered with companies that wanted to use in-lab tasks to assess their products or to, you know, to test out their products. So we've done some efficacy testing or we've used in-lab task metrics as outcome metrics for efficacy testing of nutritional supplements in partnership with a CRO. And we've built tasks for Kernel, which is a neuroscience-based company. So yeah, we're using in-lab tasks for other kinds of research.

Resources and the R community

Can I ask you if there are any resources or podcasts or books that you've liked recently that you've, you'd recommend to us? Oh my gosh. I, related to data science? It doesn't have to be, it could be, it could be more general. I don't know. I think actually I have a, a problem with that in that, like, I feel like I bookmark so many things, you know, I think whether it's Twitter or LinkedIn, you come across things that look like, oh, this will be really useful to read or, you know, useful video to watch. And there's almost like too many of them.

I mean, I definitely rely heavily on the, you know, the, our community, whether it's blog posts or, you know, all of the material that's free that other people who use R create, it's, you know, one of the reasons that, that I stick with R.

That's awesome. I, I see that Ian Moore is on from Appsilon and Ian, I wanted to give you a moment to share about the Shiny Conference if you wanted to jump in. Hey, Rachel. Sorry, I didn't expect you to be here. Hey, Rachel. Sorry, I didn't expect this at all. Oh, sorry. I just really just put you on the spot, but I saw you were still there. If you haven't heard about the Shiny Conference yet, it's happening at the end of the month and we have an incredible lineup. We're super excited about it. This is going to be the first Appsilon Shiny Conference. So there's, you know, some things we're working out, but the general time zone, I believe is from 5am to I think 1pm Eastern time. But most of the main content is going to be happening later in the day.

Well, thank you so much, Jenny, for joining us today and for sharing all your insights and about state space. It's great to, great to chat with you. And I appreciate you mentioning the need to have this platform to help each other review resumes too. So hopefully we can. Yeah, yeah, I'm really excited to see what happens with that. So thank you. I can see a few people already filled out the form just from the call today, so it's awesome.