Data Science Needs You with Andrea Jones Rooy, Ph D

Read the Transcript

Allison: Welcome back to the Deliberate Leaders podcast. I am your host and Executive Coach Allison Dunn. I have a fantastic guest with us today who is going to be diving deep into data science needs you as our episode topic we have with us Andrea Jones, right? She is a PhD, is a data and social scientist, a science educator, stand up comedian, and circus performer, which I think is super cool. Andrea is also a research consultant and keynote speaker for Global Fortune 500 companies and tech companies on how to thoughtfully integrate data science into achieving their goals, especially in the people analytics space. Andrea, thank you so much for joining us today.

Andrea: Thank you so much for having me. And for the thorough introduction. Normally, I don’t you know, delve into the comedy circus stuff. So that’s always fun to talk about.

Allison: I think that that’s I want to learn more. So ask you a question about that. Excellent. I love to kick these off with a deliberate conversation, what would be your number one leadership tip for our listeners?

Andrea: Well, I’m going to stay in on brand. So not comedy and circus though I recommend both, I would say to leaders that data is your friend. And data is something to actually think about getting your hands dirty. And working with directly even if you think you don’t have time, it’s helpful to look at a spreadsheet and just sort of see what the numbers are. And to listen to your data scientists who tell you things that are easy to do or hard to do. Or if they say this might not be worth the effort. They probably know what they’re talking about. And so data scientists, if they do say no to something, they’re not doing it out of laziness, they’re doing it because they have ideas about what might be more fruitful. So I guess more conversation with your data and with your data scientists. Okay, I’m defensive, I’m going to criticize data scientists to Don’t worry.

Allison: So I think that’s a fantastic tip. And I think it leads to a lot of questions for me, as I’m sure it does with a lot of other businesses who may not be using data to their advantage. So let’s just kind of go baseline. So what is data science?

Andrea: I’m so excited you asked, I actually, we just kicked off a new semester at NYU, where I teach data science. And so I’ve literally been expounding on this very topic for hours in front of maybe unimpressed 18 year olds, so I’ll keep it snappy. You’re here. But data science is basically using data and the scientific method to better understand the world. And a lot of people get really excited about the data side of it, I need more data, what does the data say? What can the data tell me the data has all the answers, but I really think that we can do more, oh, my gosh, I have to let someone in I’m sorry. Oh, my gosh, ah.

Data science is about using data and science to better understand the world. And we tend to get really focused on the data side of things. And there’s new data in the world data is full of lots and lots of insights. But it’s the science side of things that helps us actually discover new things and better understand the world. And so science is literally what we all learned, you know, in elementary, middle high school, when we did the scientific method, right, start with a theory, have a hypothesis, and explore that hypothesis in the data. And so I think of data science as very people driven. So we have ideas about what’s interesting, we have questions about the world, we have business challenges we want to understand or improve or predict. And we ask those questions, and then turn to the data for information, as opposed to where a lot of people I think get hung up is they focus on data driven solutions, and you just kind of look at dashboards and don’t know what to make of it. Right. And so I say, data science is about doing science, understanding the world and using data to help us better understand that world.

Allison: Thank you for kind of the overview, is there I mean, I realized we can solve for so many different things. So I guess my question would be is, is there a common theme that every business really should be desiring to look at the data to support their decisions? What would the question be? Or what would we be solving for?

Andrea: Hmm, that’s a great question. And yeah, I mean, there’s definitely no universal answer, except for and let me maybe tell me if this is too abstract, but actually, you know, the question we’re almost always trying to answer is why write a question of causality. So why do people want to work at my company? Why don’t they want to work at my company, why our sales going up? Why our sales not going up? I think in data science, we get so excited.

A lot of people use data for predictions and predictive analytics. And you know, if I A be testing on a subject line for an email to my customers and which and that, and that’s all great. And there’s a lot of really interesting models that do a pretty good job of predicting what’s going to cause, you know, engagement online to go up and that sort of thing. But the more powerful question, I think, and this is the science bias of mine, is the why, right? Why are more people coming to work here? Why are people leaving? Why are people responding to this subject line? Not just are they and should I use more emojis in my subject line, whatever it is that we’re working on, for, for marketing, but why right, what is it about that?

And so I think asking, why not just what or how, or to what extent, you know, we can go a little bit deeper than the descriptions, we can kind of start to probe causality. And I think that makes us ultimately less reliant on data. Because we understand Oh, it’s people like, when our marketing messages have a personality, it’s not really just that, oh, I have to put an emoji every time. Can you tell? I’ve been reading about marketing. I can. Okay, great. Should I be using emojis in my marketing lines? You know, the research suggests that you should, but my argument is that once everybody does that, then that’s no longer going to be helpful. And I don’t know about you, but my inbox increasingly is filled with emojis. And I kind of glanced over them. So fair enough. Yeah. But look, I like an emoji. So why not?

I think it’s if it’s appropriate. I do think it stands out for sure. Yeah. Okay. So solving for y and really asking like having good y questions. Yes. When you when you’re seeking data, understanding the why behind it, right. And making sure it’s already making sure that we’re not making assumptions about the Y. So as you said, in the opening, I do a lot of space work in this people analytics space, you take like an engagement survey that most of us have interacted with, whether we’re taking it or analyzing it or making sense of it.

A lot of times, we’ll say, oh, engagement went up, or engagement went down. And then the leader will say, Oh, it’s because of this policy. It’s because we need more recharge dates, it’s because but we haven’t tested that part. That’s your hunch. And what you need to do from there is say, Ooh, could it be recharged days? Could it be burnout? If so, let me go and probe and work with the data to see if I can get answers to that question before I leap to some conclusion, right? There’s a million different ways and we may not know the actual why. Okay, so I’m in the interim, you kind of lead into my next question.

Allison: So you say that you specialize in the people analytics, analytics, space engagement being a slice of that pie. What else does that mean?

Andrea: Right? I’m so glad you asked. I could talk about this until the end of time. So certainly engagement worker experience any kind of quantitative information we collect about what the employee journey is, like, what people’s pain points are, what people enjoy about working at the company, so anything about that, but it also falls under, you know, includes hiring, so using data to find, you know, candidates who are a great match for the company, figuring out who to reach out to in the first place, all of those things are things that we can use data for, and we see it being used. And it doesn’t mean that we’re all using it correctly, right?

There’s a lot of dangerous biases that come out and say, natural language processing of resumes and things like that. But hiring and recruitment for sure. But it also comes down to performance evaluation quite a bit, which I spend a lot of time working with companies on where we all want data on how our employees are performing, we want to make sure that we’re recognizing the heavy air quotes, top performers, right, I put it in heavy air quotes, because most companies don’t really go far enough, I think in in, in defining what that even means. And we’re all obsessed with turning our performance into numbers. Am I doing better? And is this team productive? What can we do you know, just the addition of this person or this process, improve productivity for our team? And so turning all of that thoughtfully, ethically fairly into numbers falls under People Analytics as well. Okay, fantastic. That’s a big pie. It’s a big bite. Yes. Sometimes I’m like, I should branch out. And then I’m like, No, I’ve way too much to think about already. From the perspective of you know, me and just to conceptually have numerical, and in decision driven answers around any of one of those things is incredibly powerful thing.

Allison: Yes. Who, inside of a business should be assisting with that type of data? Or is this an extra thing that you can get guidance on?

Andrea: Right? Well, the conflict of interest answer is you need an external person like me, but we don’t. You don’t actually so I’m not much of a marketer. As you can see, I’ll put an emoji in my subject line, but that’s sort of the end of it. You can do this all internally. And there’s kind of two big stakeholders that I would encourage you to either you know, tasked with this work or think about upskilling, or recruiting or whatever works best for your environment. And those two are a mix of someone who is comfortable working with data and the analytics side of things. And this doesn’t have to be someone with a formal data science degree. Those are only relatively recent. Anyone with a social science, statistics, mathematics, pewter, science, that should be fine, right? So someone who can work with the numbers and create visualizations and is comfortable making sense of spreadsheets, right?

That’s one side of it. And then the other side of it is subject matter experts. And this is an area that I think a lot of people don’t think about when they think about data science, they think about the numbers and the Mr. Robot and the hackers at a laptop all night long. And that part’s there, and it’s cool. But we also need people who know the business, who know the people who will know what recent policies have been implemented, who can speak to some of those wise, so we see your Trend engagement is going up, who’s on the ground, who’s working with employees, maybe an HR business partner, maybe you’re someone who works in the C suite, or you know, the head of people, I’m focused on people who has their finger on the pulse of the company says I suspect, the reason it’s going up is because we’ve changed this thing, or I suspect it’s gone up now because of this event, but it may not stay up. So someone who knows the business, sometimes that’s the same person. And maybe in the long run, that will be the same person eventually. But right now, I think because data science is so new, usually, I see the most effective setup being a collaboration between like the analytics science person, and the subject matter expert, by the way, this applies outside of people analytics, right? It would apply to marketing, like I can look at marketing data, but my goodness, I need a marketing expert to tell me, what might be an interesting question to ask what would be a trend that would be worth further probing? And so on? Okay, good. Thank you for kind of like the balance of you do need the institutional like pulse part of it totally.

Andrea: I think, too often we rely on one, yes. Without the other, right. And then we make a bunch of assumptions or intuition without, we’re just data but not really, to understand the impact. And I’ve seen companies really, you know, very well meaning working very hard, but really be led astray where they have like a data science team that does a whole bunch of predictive analyses on one particular example was a company wanted to predict why people would leave the company. And so they did all this work. And they figured out Oh, it’s related to salary, everyone’s like, okay, I guess we’ll pay everyone more, which is a good outcome for the employees, right. But it turned out that way, later on, they went back and debriefed with the actual business experts and the leaders in the company. And it turned out that the model had left out bonuses, which are a huge part of employee compensation. So the entire prediction was incorrect, not because the model was wrong, or the data scientist wrong, but they just didn’t know the full picture of the business. So that’s a very stark example. But that kind of thing happens all the time.

Okay. And honestly, like a holistic, like, you have to have both working together, right? That’s right. I’m also glossing over a really hard part, which is collaboration between those two worlds. So this is where I’ll be critical of data scientists, we need to work harder to make sure that other people understand what we’re doing and what’s in the data and be transparent about the limitations of our analyses. I think there’s a lot of temptation for data scientists to start throwing out jargon and machine learning and neural nets and metal bla bla. And if no one knows what you’re talking about, you’re not actually helping anyone and no one’s better understanding the world. So that’s my criticism for data scientists, our job to be understood.

Allison: That’s, that’s a good Aha, you know, reflection, I appreciate that. You’re, clearly you’re incredibly passionate about data science in general, what’s the best part of data science?

Andrea: Oh, I’m getting like chills just thinking about it. The best part is that we get to study almost everything, right? Like you have to have that humility that I’ve been talking about. And I have to remind myself of that, which is that just because I can make sense of the numbers, or put the numbers in a pretty picture that looks interesting, doesn’t mean I understand the underlying subject matter or the area. But it does mean that I you know, especially because I love working across different disciplines is I get to tackle lots of different types of problems. And so I do most of my work and people analytics, as we said, that’s already a lot of different types of projects. But every now and again, I also do partner with other parts of the business where we talk sales, or we talk, finance and budgeting or other things. So there’s all kinds of different questions that we could ask.

Allison: Okay, cool. What’s the hardest part of data science?

Andrea: I would say it’s, and this is really a more of a personal assessment. It’s managing is it’s a new field. So a lot of people don’t know what it is. And I’m not convinced that I have nailed it. Writing what it is yet. And what that translates to is, is I need to work very hard to be clear about what’s possible and what’s not with data. And lots of times, it can be very in the weeds where you say, well, this analysis you’re describing would be super interesting. But in order to do that, we would need the data to be structured in a totally different way. And it would actually take like 40 hours of non stop just data wrangling boring stuff, to get that in a position where it would make sense. And so I as the external consultant to charge a lot more because it’s really, but it’s not obvious why that’s hard, because very few people have experience with that.

So it’s more like explaining what’s, what’s possible, and what’s hard and what’s easy. And data science is not always obvious. And I think that’s true in any, any field where you’re, you’re used to something, you say, Oh, this is actually a really hard thing, even though it seems like it might not be or this is actually trivial. But everyone’s like, Wow, that’s amazing, right? The same is true in circus. So bring it back. Yes, some tricks look awesome, but are very easy. And some tricks look not that impressive, but are really hard and communicating that to the other, you know, to audiences is kind of the key.

Allison: I am I don’t I don’t really no, if I’m asking like a question. That’s going to make sense. But okay, so my, my son has a degree in geography, geography information systems. So GIS, yeah, you’re a data scientist. And then I work with companies that have like business intelligence, the difference? Why are they so different?

Andrea: Wonderful. So I’m going to make a lot of people furious by saying that there’s a ton of overlap in all of those. And at the heart, it’s all about turning the world into, you know, data or information that we can, you know, condensing the world or distilling the world into something that we can make sense of a simplified version. So my I do not do GIS myself. But I could argue that GIS is a technique that many data scientists practice, but it also lives outside of, of data science, and is its own, you know, method that you could apply to say environmental science and it kind of can exist in the subject matter expertise area, I tend not to get too fussed about the distinctions, though, I think that’s changing.

Now, there’s like data scientists and data engineers, and never the two shall meet, right. But generally speaking, I say if you are working with data in a scientific manner, you’re doing data science, even if you don’t call yourself a data scientist. And sometimes data scientists live very much in just one field. So my original field was political science. I have political science colleagues, who exclusively work with politics data. And their research really is substantively about predicting elections and turnout and things like that. But the methods that they’re developing are perfectly applicable to some other field. And people in political science use GIS and people in all kinds of different fields. So it’s really more, if you’re practicing the methods that fall under Data Science, then I think you’re doing data science. Okay. So I appreciate that there’s a lot of overlap. And essentially, it’s all data.

Andrea: In some way, yeah. According to Andrea, according to me, yes, yes. And some people might get faster and say, Well, Mike, you’re describing maybe as statistics. And I guess, you know, we’re starting to see some argument where it’s like, you have to have some level of like, computational intensity to count as data science like, like machine learning is data science in a way that, you know, correlation is not, but I don’t think that distinction is particularly useful. Right. Okay. All right. Fair enough.

Allison: You have a lot of varying interests, is what I would say. And so the first one I just want to ask you a little bit more about so you have a podcast is called majoring in everything. And that you’re working on a book that is focusing on why focusing on one ever rated. Yeah. Tell us more.

Andrea: Yes. So it’s, it’s the book is, is very much in progress. And it is, I do constantly think like, what should this book be about, and it’s kind of about not focusing is the story. So my entire life, I won’t get into a huge sob story. But my entire life, I’ve held up an ideal that the ultimate success is someone who finds the thing that they’re passionate about, and then pursues that thing at all costs. And I don’t know if it’s the time I grew up, or the media I was assuming or where I got that idea, but I have it and a lot of people I know, frankly, of different ages also have that idea. And it’s taken me a very long time to say what if the fact that I am naturally interested in different things is an asset and not something that I should get mad at myself about?

I spent a lot of my life because I am interested in comedy. I’m interested in movement dance circus, I do trapeze and fire, and of course, this data science but I also love teaching And I also love working in industry. And I used to say, oh, pick something, you’re wasting your time. And only in the last maybe decade, maybe less, I’ve started to say, actually, what if they’re all helpful to me. And the comedy helps the science and the science helps the comedy and all of it helps the circus and so on. And so the book is exploring, and the podcasts are both exploring, talking to other people who have similar kind of multifaceted multi dimensional paths, figuring out the unexpected ways that those connect and overlap, and maybe change the narrative from, you know, success being like at the top of the mountain to being like, it’s actually pretty cool if you’re kind of involved in three different mountains. I’m working out the analogy still. But that’s the idea.

Allison: Okay, fantastic. I love it. Andrea, I just am thrilled to kind of provide a better understanding of how businesses and leaders can make better sense and how to make better sense of data and the science behind it. And what is the best way for people to connect and or follow you?

Andrea: Yes, so you can find me on all the social medias and all of that stuff. At Jones ROI, J o n e. S. Roy, it’s just my last name with no hyphen, and Jones roi.com. And from there you can you know, if you message me on any or Andrea Jones or on LinkedIn, and I can tell you that you’re not a bot, we can email.

Allison: Fantastic. Andrew, thank you so much for your time today and keep on keep on cranking on that data. Thank you so much same to you. Thank you.