(energetic music) >> Announcer: Live, from Stanford University in Palo Alto, California, it's The Cube, covering Women in Data Science Conference 2018. Brought to you by Stanford. >> Welcome back to The Cube. We are live at Stanford University at the third annual Women in Data Science Conference. I am Lisa Martin. We've had a really exciting day so far, talking with a lot of female leaders in different parts of STEM fields. And I'm excited to be joined by my next guest, who is a speaker at this year's WIDS 2018 event, Daniela Witten, the Associate Professor of Statistics and Biostatistics at the University of Washington. Daniela, thanks so much for stopping by The Cube. >> Oh, thanks so much for the invitation. >> So here we are at Stanford University. You spent quite a lot of time here. You've got three degrees from Stanford, so it's kind of like coming back home? >> Yeah, I've spent from 2001 to 2010 here. I started with a bachelor's degree in math and biology, and then I did a master's, and finally a PhD in statistics. >> And so now you're up at the University of Washington. Tell us about that. What is your focus there? >> Yeah, so my work is in statistical machine learning, with applications to large scale data coming out of biology. And so the idea is that in the last ten or 20 years, the field of biology has been totally transformed by new technologies that make it possible to measure a person's DNA sequence, or to see the activity in their brain. Really, all different types of measurements that would have been unthinkable just a few years ago. But unfortunately, we don't yet know really how to make sense of these data statistically. So there's a pretty big gap between the data that we're collecting, or rather, the data that biologists are collecting, and then the scientific conclusions that we can draw from these data. So my work focuses on trying to bridge this gap by developing statistical methods that we can use to make sense of this large scale data. >> That sounds exciting. So, WIDS, this is the third year, and they have grown this event remarkably quickly. So, we had Margot Garritsen on the program a little bit earlier, and she had shared 177 regional WIDS events going on today, this week, in 53 countries. And they're expecting to reach 100,000 people. So, for you, as a speaker, what is it that attracted you to participate in the WIDS movement, and share your topic, which we'll get to in a second, what was it that sort of attracted you to that? >> Well, first of all, it's an honor to be invited to participate in this event, which, as you mentioned, is getting live streamed and so many people are watching. But what's really special for me, of course, as a woman, is that there's so many conferences out there that I speak at, and the vast majority have a couple of female speakers, and it's not because there's a lack of talent. There are plenty of very qualified women who could be speaking at these conferences. But often, the conference organizers just don't think of women right away, or maybe add a couple women as an afterthought to their speaker lineups. And so it's really wonderful to be part of a conference where all of the speakers are women, and so we can really see the broad ways in which women are contributing to data science, both in and out of industry. >> And one of the things that Margot shared was, she had this idea with her co-founders only three years ago in 2015, and they got from concept to their first event in six months. >> Daniela: Women know how to get things done. >> We do, don't we? (laughs) But also what it showed, and even in 2015, and we still have this problem in 2018, is there's a massive demand for this. >> Yeah. >> The statistics, speaking of statistics, the numbers show very few women that are getting degrees in STEM subjects are actually working in their field. I just saw this morning, it's really cool, interactive infographic that someone shared with me on Twitter, thank you very much, that showed that 20 percent of females get degrees in engineering, but only 11 percent of them are working in engineering. And you think, "How have we gone backwards in the last 30 years?" But at least now we've got this movement, this phenomenon that is WIDS to start, even from an awareness perspective, of showing we don't have a lot of thought diversity. We have a great opportunity to increase that, and you've got a great platform in order to share your story. >> Yeah. Well, I think that you raise a good point though, as, even though the number of women majoring in STEM fields, at least in some areas of STEM has increased, the number of women making it higher up in the STEM ladder hasn't, for the most part. And one reason for this is possibly the lack of female role models. So being able to attend a conference like this, for young women who are interested in developing their career in STEM, I'm sure is really inspirational and a great opportunity. So it's wonderful for Margot and the other organizers to have put this together. >> It is. Even on the recruiting side, some of the things that still surprise me are when some, whether it's universities or companies that are going to universities to recruit for STEM roles, they're still bringing mostly men. And if there are females at the events, they're, often times they're handing out swag, they're doing more event coordination, which is great. I'm a marketer. There's a lot of females in marketing. But it still shows the need to start from a visibility standpoint and a messaging standpoint alone. They've got to flip this. >> I completely agree with that, but it also works the other way. So, often a company or an academic department might have a few women in a particular role, and those women get asked to do everything. Because they'll say, "Oh, we're going to Stanford to recruit. We need a woman there. We're having some event, and we don't want it to look totally non-diverse, so we need a woman there too." And the small number of women in STEM get asked to do a lot of things that the men don't get asked to do, and this can also be really problematic. Even though the intent is good, to clearly showcase the fact that there's diversity in STEM and in academia, the end outcome can actually be hurtful to the women involved who are being asked to do more than their fair share. So we need to find a way to balance this. >> Right. That balance is key. So what I want to kind of pivot on next is, just looking at the field of data science, it's so interesting because it's very, I like 'cause it's horizontal. We just had a guest on from Uber, and we talk to on The Cube, people in many different industries, from big tech to baseball teams and things like that. And what it really shows, though, is, there's blurred lines, or maybe even lines that have evaporated between demarcated career A, B, C, D. And data science is so pervasive that it's impacting, people that are working in it, like yourself, have the ability to impact every sector, policy changes, things like that. Do you think that that message is out there enough? That the next generation understands how much impact they can make in data science? >> I think there is a lot of excitement from young people about data science. At U-dub, we have a statistics major, and it's really grown a lot in popularity in the last few years. We have a new master's degree in data science that just was started around the same time that WIDS was started, and we had 800 applicants this year. >> Wow. >> For a single masters program. Truly incredible. But I think that there's an element of it that also maybe people don't realize. So data science, there's a technical skill set that comes with it, and people are studying undergrad in statistics, and getting master's in data science in order to get that technical skill set. But there's also a non-technical skill set that's incredibly important, because data science isn't done in a vacuum. It's done within the context of interdisciplinary teams with team members from all different areas. So, for example, in my work, I work with biologists. Your previous guest from Uber, I'm sure is working with engineers and all different areas of the company. And in order to be successful in data science, you need to really not only have technical skills, but also the ability to work as a team player and to communicate your ideas. >> Yeah, you're right. Balancing those technical skills with, what some might call soft skills, empathy, collaboration, the ability to communicate, seems to be, we talked about balance earlier, a scale-wise. Would you say they're pretty equivalent, in terms of really, that would give somebody a great foundation as a data scientist? >> I would say that having both of those skill sets would give you a good foundation, yes. The extent to which either one is needed probably depends on the details of your job. >> True. So, I want to talk a little bit more about your background. Something that caught my eye was that your work has been featured in popular media. Forbes, three times, and Elle magazine, which of course, I thought, "What? I've got to talk to you about that!" Tell me a little bit about the opportunities that you've had in Forbes and in Elle magazine to share your story and to be a mentor. >> Yeah. Well, I've just been lucky to be getting involved in the field of statistics at a time when statistics is really growing in importance and interest. So the joke is, that ten years ago, if you went to a cocktail party, and you said that you were a statistician, then nobody would want to talk to you. (Lisa laughs) And now, if you go to a cocktail party and you say you're a statistician, everyone wants to know more and find out if you know of any job openings for them. >> Lisa: That's pretty cool! >> Yeah. So it's a really great time to be doing this kind of work. And there's really an increased appreciation for the fact that it's not enough to have access to a lot of data, but we really need the technical skills to make sense of that data. >> Right. So share with us a little bit about the session that you're doing here: More Data, More Statistical Problems. Tell us a little bit about that and maybe some of the three, what are the three key takeaways that the audience was hearing from you? >> Yeah. So I think the first real takeaway is, sometimes there's a feeling that, when we have a lot of data, we don't really need a deep understanding of statistics, we just need to know how to do machine learning, or how to develop a black box predictor. And so, the first point that I wanted to make is that that's not really right. Actually, the more data you have, often the more opportunity there is for your analysis to go awry, if you don't really have the solid foundations. Another point that I wanted to make is that there's been a lot of excitement about the promise of biology. So, a lot of my work has biomedical applications, and people have been hoping for many years that the new technologies that have come out in recent years in biology, would lead to improve understanding of human health and improve treatment of disease. And, it turns out, that it hasn't, at least not yet. We've got the data, but what we don't know how to do is how to analyze it yet. And so, the real gap between the data that we have and achieving its promise is actually a statistical gap. So there's a lot of opportunity for statisticians to help bridge that gap, in order to improve human health. And finally, the last point that I want to make is that a lot of these issues are really subtle. So we can try to just swing a hammer at our data and hope to get something out of it, but often there's subtle statistical issues that we need to think about, that could very much affect our results. And keeping in mind sort of the effects of our models, and some of these subtle statistical issues is very important. >> So, in terms of your team at University of Washington, or your classes that you teach, you work with undergrads. >> Yeah, I teach undergrads and PhD students, and I work mostly with PhD students. And I've just been lucky to work with incredibly talented students. I did my PhD here at Stanford, and I had a great advisor and really wonderful mentoring from my advisor and from the other faculty in the department. And so it's really great to have the opportunity now, in turn, to mentor grad students at University of Washington. >> What are some of the things that you help them with? Is it, we talk about inspiring women to get into the field, but, as you prepare these grad students to finish their master's or PhD's, and then go out either into academia or in industry, what are some of the other elements that you think is important for them to understand in terms of learning how to be assertive, or make their points in a respectful, professional way? Is that part of what you help them understand and achieve? >> That's definitely part of it. I would say another thing that I try to teach them, so everyone who I work with, all my students, they're incredibly strong technically, because you don't get into a top PhD program in statistics or biostatistics if you're not technically very strong, so what I try to help my students do is figure out not just how to solve problems, because they can solve any problem they set their mind to, but actually how to identify the problems that are likely to be high impact. Because there's so many problems out there that you can try to solve statistically, and, of course, we should all be focusing our efforts on the ones that are likely to have a really big impact on society, or on health, or whatever it is that we're trying to influence. >> Last question for you. If you look back to your education to now, what advice would you give your younger self? >> Gosh, that's a really great question. I think that I'm happy with many of the career decisions I've made. For example, getting a PhD in statistics, I think is a great career move. But, at the same time, maybe I would tell a younger version of me to take more risks, and not be so worried about meeting every requirement on time, and instead, expanding a little bit, taking more courses in other areas, and really broadening instead of just deepening my skill set. >> We've heard that sentiment echoed a number of times today, and one of the themes that I'm hearing a lot is don't be afraid to get out of your comfort zone. And it's so hard for us when we're in it, when we're younger, 'cause you don't know that, you don't have any experience there. But it's something that I always appreciate hearing from the women who've kind of led the way for those of us and then, the next generation, is, don't be afraid to get comfortably uncomfortable and as you said, take risks. It's not a bad thing, right? Well, Daniela, thanks so much for carving out some time to visit us on The Cube, and we're happy to have given you the opportunity to reach an even bigger audience with your message, and we wish you continued success at U-dub. >> Oh, thanks so much. >> We want to thank you for watching. I'm Lisa Martin live with The Cube at WIDS 2018 from Stanford University. Stick around, I'll be back with my next guest after a short break. (energetic music)