 Live from Stanford University in Palo Alto, California. It's theCUBE, covering Women in Data Science Conference 2018. Brought to you by Stanford. Welcome back to theCUBE, I'm Lisa Martin at Stanford University. We're here for the third annual Women in Data Science Conference. Hashtag WIDS 2018, check it out, be part of the conversation. WIDS is in its third year, but it's aiming to reach about 100,000 people this week alone. There's 177 regional WIDS events in 53 countries. This event here, the main event at Stanford, features keynote, technical vision talks, a career panel, and we're excited to be joined next by Dr. Natalie Henri-Riche. I did that in French. Yes. Who is a researcher at Microsoft. And Natalie, first of all, welcome to theCUBE. Thank you, I'm really thrilled to be here. Yeah, you gave a technical vision talk on data visualization and data-driven storytelling. Share with our audience some of the key messages that the WIDS audience heard from you earlier today. Well, I guess I gave two main messages. The first one is data visualization has two superpowers. Superpowers? Superpowers. Tell me, girl. The first one is enable you to kind of think about your data in a new way. So just kind of form hypothesis and answer questions you didn't even know you had by your data. So that's the first one. And the second superpower is it's really useful to communicate information and communicate with a large audience. Visualization helps you kind of convey your point with data to back it up. So that's kind of the short one minute. I love that super, super hero, super power. So WIDS is, as I mentioned in the intro, in its third year and reaching, it's grown dramatically in such a short period of time. This is your first WIDS. And your first WIDS, you are a speaker. What was it that attracted you to WIDS? And you went, yes, I want to give some of my time to this and come down from Seattle. Well, so I'm French originally. And my studies, I did engineering school and was one of three out of 300 men, right? Wow. So I was requested a lot for women in computer science, in engineering. And so I actually really like it. Just meeting all of those people, talking about trying to bring more women in. Part of the job I'm doing is very creative. So trying to come up with new ideas for visualization. And I think having a wide range of people just adds to the mix. And we get so many more exciting ideas. So I really want to try to have more diverse group of people I can work with and connect to. And so that's what attracted me to here. Excellent. A couple of things that you said, I've heard a number of times today. And the first one is what Daniella Witton shared, who's also a speaker that oftentimes some of the few women in tech, and you mentioned being one of three in 300, are asked to do a lot of other things. Did you find that that, okay, you're one of the few females, you articulate, you like speaking, we want you to do all these things? Yes. And I say no to that. That's a skill too. Because I have kids too, so. But yeah, I mean, it happens a lot, but I think as we go further, it's going to be less and less happening. And it's better in the end. So it's kind of a service, I see it as a service to my field and my company. But at the same time, we also get a lot of benefits from it. But that said, I tried to cut it down to a manageable level. So two hours flight from Seattle works great. Right, right, right. Another thing is that you mentioned the creativity. And I've heard that a number of times today from our guest, Margot Garrison, was on as well. Tell me about your thoughts about being in this data science role, the need for creativity. How does, why is that? You might consider like a softer skill versus the technical skills, but how important is that creativity in your job, for example? So my job is really like researchers, trying to have new ideas and innovate for Microsoft in particular. So I'm not really a data scientist, but I build the tools for a data scientist. So knowing that creativity is important because you need to kind of think out of the box what is the next generation of tools that they would need. And in turn, they need to think out of the box, kind of get more insight of the data they're collecting. So creativity is just like pervasive to this whole data science thing. And problem solving as well. So you need a lot of the left brain and a lot of the right brain, kind of both of them together. And I think that having different cultures and different genders, even different age ranges, just makes you think out of the box, that's just what's happening. Discussing with people, I was discussing with someone in the cosmology and as well that brought up a lot of different ideas in me. So to me, that's really critical part of what I'm doing every day. I like that, that kind of aligns to what one of our guests said earlier and that is thought diversity. I've never thought of thought diversity, but you bring up a good point about, it's not just about having women in the field. It's also about having diversity in terms of generations. And one of the things that's, I think, pretty unique about WIDS is, it's not just about reaching young women in their first semester at university, for example, as Maria Claude has said, that's the ideal time to really inspire. But it's also reinvigorating women who have been in academia or industry in STEM subjects for a long time. So we have multiple generations. And to your point, it's like, that diversity is important. It's not just about gender or ethnicity. It's also about those diverse perspectives that come from being from different generations. So it's funny because I was giving this talk earlier and one part of it was about timeline. And when I was researching how people draw time, well, depending on some culture, it goes from left to right, but some other culture, it's front to back, back to front, right to left. So we need to be aware of all of that. And it's so much easier to just have the people to converse with right in your office or next door to be aware of those. So that's very important, especially to big companies like Microsoft, except a lot of customers worldwide. So it's very important to just be immersed in that. Definitely. So you have been published. You've got published research and over 60 articles and leading venues in human computer interaction and information visualization. But something we chatted about off camera was very intriguing about visualization and children. Yes, tell me a little bit more about that. So I happened to have two kids, seven and four. And I'm passionate about what I'm doing and I just couldn't keep it out of their hands, right? So I was just starting seeing what does my daughter learn at school? What does she learn in kindergarten? And in fact, in kindergarten, I remember one day she brought back candies and I'm like, did you get candies from school? She's like, no, because we were doing a bar chart. I was like, what? And so I was very intrigued in what do we teach? What do your kids learn? And it was fascinating to see that from an early age they learn how to do those visualization, but they don't really learn how you can lie with them or think critically about that, that maybe you can start your bar chart at two and you would have less candy, I guess, but you could kind of convey the wrong messages. So I became passionate about this and decided we need to just improve our teaching about how we can represent data and how we can also misrepresent it in the hope that for the next generation to come, they'll be able to look at a chart and think critically about it, whether or not it tells the right story with the right data, kind of beyond just a picture is worth a thousand words and I'm not going to think about it. Yeah. So it's kind of my personal effort that I try to move myself forward. That was so important, that having that passion. I think that's one of the things that seems to be inherent about WIDS. Even yesterday, seeing the, on the Twitter stream, WIDS New Zealand starting in five minutes and it's been really focused on being so kind of inclusive just sort of naturally. And one of the things that I learned in some of my prep for the show is the bias that is still there in data interpretation. And you kind of talked about that and I never really thought about it in that way but if a particular group of people is looking at a data set and thinking it says this and no other opinions, perspectives, thoughts are able to be incorporated to go, well maybe it says this. Then we're limiting ourselves in terms of one, the potential that the data has to help a new, a business create a new business model but also we're limiting our perspectives on making a massive social impact with data. Yeah, what I found very interesting is visualization, often people think about it at the end of the spectrum, like I've collected my data, I analyze it and now I need a pretty picture to kind of explain what I found. But the most powerful use of visualization I think comes early on where you actually just collected your data and you look at it before you run any statistical test. And then I did that not long ago with French air traffic data in the whole lens, I put them in and I saw the little airplanes moving around. Then what we saw is one airplane doing loops like this. I was like, what is this going on, right? And it was just a drone doing like tests, right? But somehow it got looped into that data set. And so by looking at your data early on, you can detect what's wrong with the data. So then when you actually run your statistical tests and your analysis, you better reflect, what was that data in the first place? What could go wrong there? So I think inserting visualization early on is also critical to understand what we can really know and do and ask about the data in the first place. So it's kind of like watching this story unfold rather than going, we've done all this analysis, here's the picture, the story is this. The story is you're sort of turning it sort of page by page it sounds like and watching it and interpreting it as it's unfolding. And rethinking what you collected in the first place. Is that the right data? You collected to answer the question you wanted to ask, is it a good match or not? And then rethink that, collect new data or the missing one and then go on with your analysis. So I think to me it's really a thinking tool. It also sounds like another, we talked about like the technical skills that obviously that a computer scientist, data scientist needs to have, but there's other skills. Empathy, communication, collaboration. Sounds like also there needs to be an ideal kind of skill set has to include open-mindedness. Tell me a little bit about some of your experiences there and not being married to the data must say this. So if it doesn't, I'm not going to look anywhere else. Where is open-mindedness in terms of being a critical skill set that needs to come to the field? Yeah, I mean that is totally a really critical point. Think already when you're collecting the data, especially as a scientist when I run experiment, I kind of know what I want to find. And sometimes you don't find it and you need to kind of embrace it. But it's hard to have because sometimes it's like those unconscious bias you have. Like you're not really necessarily controlling them and just the way you collected the data in the first place maybe just skewed your results. So it's very important to kind of think ahead of time of all of those bias you could have and think about all of what could go wrong. And often the scientific process is actually that, trying to think about all of the stuff that could go wrong and then check whether or not they're wrong. And we're trying to infuse that a little bit over Microsoft as well, kind of the data we collect. Can we analyze them? Can we have teams of people who really think is that the right data? Are we collecting like worldwide for example or are we just collecting from the US? So there's a lot of those kind of ethical and bias kind of training and effort, try to remove that the maximum from our work. And I think it's across the entire world. I think we solve this data collection everywhere. We kind of have to do that very consciously. I think two things kind of speak to me that out of what you just said that we've heard a number of times today. One that failure, and I don't mean to say that, failure is not a bad thing. No, that's how you know it. Exactly. And growth. Exactly. In many ways it's not a bad effort. This is how everybody that's successful got to wherever they are. But it's also been embracing as you said the word, embracing the fact that you might be bringing bias into this and you have to be okay with, maybe this is the wrong data set. If you consider that a failure, consider it to your point a growth opportunity. That is one of the themes that we've heard today and you kind of elaborate on that. The second one is be okay getting uncomfortable, get out of that comfort zone. Consciously uncomfortable because when you're able to do that, the possibilities are limitless. Yes. And that's what I try to do every day because I try to push all of the software that we're doing and Microsoft is so big and all of those software are like so there. So trying to come up with new ideas like so many are failures, they won't make money or they don't actually work for this population. So most of my work is failure. But hey, one success when in a while and I'm happy about it. Exactly. But it's just charting that course to getting to the, this is the pot of gold at the end of the rainbow. Well Natalie, thank you so much for taking some time to talk with us on theCUBE and sharing your story. Congratulations on being a speaker, your first WIDS and we look forward to seeing you back next year. Thank you very much. We want to thank you for watching theCUBE. I'm Lisa Martin live from WIDS 2018 at Stanford University. Stick around, I'll be back with my next guest after a short break.