Episode Transcript
Speaker 1 00:00:24 Today joining us for our podcast is Haley ca. Haley has a background in chemical engineering and served as lead data scientist at decode m a market research firm where she founded and built their data science practice and developed analytical tools with NLP for companies like Lyft, Airbnb, WeWork, and MI more. Haley is now New York Director of Knowledge Society. Haley, thank you so much for joining us.
Speaker 2 00:00:53 Hi. Yeah, thanks for having me. I'm excited. So,
Speaker 1 00:00:56 Hailey, when we say that the Knowledge Society solves some of the most important problems in society today, what are we talking about? What are those issues and what are the solutions that you all are coming up with?
Speaker 2 00:01:07 Right. Yeah, so, so the Knowledge Society is actually a, it's a STEM accelerator program for high school students. Um, and we have it in a bunch of cities all across America and Canada, um, in a virtual global program as well. So, um, you know, we train the kids on emerging technologies like quantum computing, ai gene editing, uh, nanotech, blockchain. The list kind of goes on and, you know, they get practice, uh, and we coach them on how to use these technologies to actually solve some of the world's biggest problems. Um, and beyond that, you know, we currently, right now we're doing, uh, we're working with the un uh, the students get to help figure out how to increase, um, uh, women and girls digital or inclusion in the digital economy. So they're working with the UN on that project right now too. So, kind of a wide range of skills, uh, that the students get from the Knowledge Society. Um, but really it's training the next generation of leaders
Speaker 1 00:02:01 And clearly they are brilliant at the sorting point as as young people. So, great, great opportunity. Wondering what is the role that Data science plays in your work with Knowledge Society?
Speaker 2 00:02:13 Yeah, so right now it doesn't actually play all that big of a role. I've actually kind of moved in terms of managing students and tracking data. Um, I've started using no code tools, so things like Airtable, um, just because that's more user friendly for other people on the team as well. Um, so you know what I do though, you know, I leverage my data science background more, more or less to figure out, okay, what is working, um, in terms of coaching these students on, on being able to solve the world's biggest problems and what isn't, right. So I think I would say I use more of my analytical background for that and not so much, I don't use like, you know, machine learning models for that.
Speaker 1 00:02:54 So when we talk about finding synergies in seemingly strange places, I'm thinking chemistry and then learning and then data science. There are so many different things that you focused on in your career. What is underlying all of these different characteristics, all of these different disciplines that you've worked in, and what skillset do you think has been most important?
Speaker 2 00:03:19 Mm-hmm. Yeah, that's a good question. Um, so I mean, I think the underlying theme for me personally is really, it's about impact. So it's about how can I maximize my positive impact on the world. So for instance, uh, with data science, you know, I was working with these companies, um, you know, kind of one-on-one and, and with the team at decode m um, but you know, and so I can make an impact, uh, that way or I can, you know, train all these young people who are all gonna go on to make an impact. And now I've just, you know, a hundred Xed my impact on the world ideally. So that's, um, and that's the reason I got into data science cuz I thought I could leverage that more to make a bigger impact on the world, uh, than chemical engineering at the time. So that's the underlying thing. And I think honestly the skill that has been most important is just having to figure it out mindset. Um, I pretty much just follow, you know, what am I curious about, what you know is important to me, and then I have confidence that I can figure pretty much anything out <laugh>, um, if you throw it my way. So I think that has allowed me to switch between such different roles and such different industries.
Speaker 1 00:04:24 And I know for many folks who are listening, people are thinking about contemplating changing careers, are e even delving deeper into their existing careers. Now, you professionally you train, you support education and learning for these young people. Are there huge differences between the learning path for a young adult or a teenager for that, for that, in that case are an adult, someone who's been in the industry for 5, 10, 15, even 20 years, what would the difference in your approach be in terms of training, uh, an adult versus mm-hmm. <affirmative> a child, how to think and how to develop themselves?
Speaker 2 00:05:03 Right. Yeah, that's a good question. I mean, so, you know, and, and full disclosure, I haven't worked with adults in the same way I've worked with these teenagers. So, you know, take this with a grain of salt, but I actually think, you know, you, you can use a pretty similar approach. So what I do, um, if you wanted to just train yourself on something, this is what I've done personally, right, is data science, for example. Um, I'd say, okay, what kind of, I start out saying, okay, what do, how do I learn data science? So almost like meta learning, right? Like how do I learn? I need to learn how I'm gonna learn, uh, you know, data science or whatever it may be, right? So I kind of make a map for that and then I figure out, okay, you know, what are kind of the resources I have at my disposal online courses, you know, maybe it's cohort based courses.
Speaker 2 00:05:47 I like those a lot because usually you can work on projects with other people. Um, and then, you know, I think about, okay, why am I learning this? Right? Um, and at the time, even though I didn't end up going, going into it data science, I really wanted to, um, work in, you know, the biotech or pharmaceutical industry in drug discovery. And so at the time I did, how I learned is I just did a bunch of projects revolving around data, um, from the pharmaceutical industry or from dr, um, you know, uh, drug delivery data, that type of thing. So I think doing projects, um, to practice, you know, the new skill, um, that you're learning, but doing projects that you actually care about, right? That you're actually working towards it, it just makes it easier to get stuff done. And, and I think it just really connects on a deeper level. Um, and then when you get discouraged, it's like, oh, but I'm still excited by this prospect. So, um, cuz learning on your own and as an adult definitely is tough. And I think it's about understanding like the process isn't gonna be perfect. Um, and, you know, thinking about who can you talk to, who can you go to for feedback and for help, um, and for some guidance as well.
Speaker 1 00:06:53 So what I'm hearing from you, Hailey, is the hands-on approach as opposed to other types, and then building a community of other like-minded folks, like-minded learners,
Speaker 2 00:07:03 Right? Right. And you can do that easily, you know, uh, digitally now, right? Like you can go, you're learning data science, you know, you can go on Reddit, you can go on, um, you know, stack overflow, right? There's so many different areas, uh, discord, right, to get help from people online. So yeah.
Speaker 1 00:07:18 With all of the disciplines and the theories that you're, you're learning about and that you're teaching to others, what excites you the most in terms of the future? So what disciplines, theories, ways of approaching things are you seeing now that you know it's really gonna catch us by surprise perhaps in later 2021 or in the coming couple of years? What excites you? Gimme a couple of things,
Speaker 2 00:07:40 Right? Yeah, no, there's a few things that excite me. Um, specifically. So I actually, so there's so many technologies, it's hard to choose one. So one that I'm really excited about is space technology. So obviously, you know, we all know about Elon Musk, SpaceX, nasa, et cetera, but there are so many other companies working, you know, essentially years ahead in the future and preparing for when, you know, when humans can actually travel to and from earth to space. When are, are we living up there, you know, what's it gonna look like in terms of, you know, uh, uh, just commercial space travel. Um, and so we have, there's a ton of companies that are working in that space, but they kind of have to work with the assumption that we're gonna get to that point, right? So I find that really interesting and I don't know if it'll take people by surprise, but, um, uh, but that's what really exciting for me.
Speaker 2 00:08:30 Um, and a lot of my students are working on that. Another exciting technology for me is quantum computing as well, and I think that's one that like maybe people know, you know, have heard of quantum computing but don't really know what it is. And you know, it's essentially going to exponentially increase, you know, our computing power. So, you know, something that'd take a normal, our laptop, you know, tens of thousands of years to do it could, a quantum computer could do in a minute, like some type of calculation. So that's really gonna speed up our ability to, you know, collect, analyze data and, um, and really to build, uh, other, other digital tools that we probably don't even know exist yet because we don't have, you know, a commercially available quantum computer. But I think that's get, that is gonna accelerate a lot of other technologies and a lot of other innovations.
Speaker 1 00:09:15 And just to dive deeper into that, who's building the quantum computers? Your students? Who, who's doing that? <laugh>?
Speaker 2 00:09:20 Yeah, so they, they're actually building it. So, um, a lot of companies you can actually essentially tap into their quantum computing server, or it's like a cloud qu quantum computer, so you can actually, you know, do some projects using their quantum computer, um, on the cloud. So that's what they're doing, so they're not actually building it. Cause it, it's interesting cuz obviously there's the software, they're using the software side of it, but the really technical challenge right now is the, the hardware. Um, and actually engineering something that, that works.
Speaker 1 00:09:48 I think it's only a matter of time before they, you know, started building them out. Your students <laugh>, they sound brilliant, right? <laugh> only matter time. In terms of speaking of your students, in terms of learning, what has been the most interesting thing that you've learned from your students? I know that many times we think that the teacher is the one always passing on the knowledge that the professor or the person head the head of the class. But what have your students taught you?
Speaker 2 00:10:12 Right. That's a really good question. Um, <laugh> there's been so, so many things. I think, um, you know, one thing that that they constantly teach me is the power of just following your own curiosity, right? Like, that can take you really, really far. Um, and, and it makes life more enjoyable. You know, a lot of these students are used to just like, you know, at this point slugging through online high school and, you know, college apps, SATs, all this stuff they have to do. So giving them the opportunity and saying, Hey, you can actually, here's a process to follow, but you can choose any technology you want. You get to explore anything you want. Um, and just seeing how much they grow so quickly, just having someone literally support them in that, like they're doing all the projects on their own, you know, I I'm not, I'm I'm not in there like helping them actually code, right?
Speaker 2 00:10:58 So they're figuring it out. Um, so that's one is following your curiosity. Um, and two, actually, you know, I think, you know, something I'm passionate about is, um, uh, you know, how young people like specifically Gen Z are dealing with their mental health because they're a lot more open about it than previous generations, but there's still, you know, a huge lack of resources and understanding, um, with their mental health. And I, and it's, it's so foundational that I see, you know, it's hard to succeed doing anything with tech, anything with high school, right? A anything, you know, in life if you don't have a really, or if you don't have at least a somewhat solid foundation in terms of what strategies do I use to just manage my emotions? So even if you aren't, you know, uh, if, even if you aren't struggling with a mental illness, so to say, like, actually how do I develop emotional intelligence? Um, and just coping strategies in general and just seeing like that's, that's kind of the linchpin for a lot of my students, and I think that's that's true for many, many people.
Speaker 1 00:11:55 Wow. So from coding to emotional intelligence, it sounds like it's a huge range, but you're doing great work, Hailey, thank you so much for joining us. Thank
Speaker 2 00:12:05 You.
Speaker 1 00:12:07 Viba Sub Hanian is currently VP of technology at Analytics Insight. Viva leads the technology team at Analytics Insight and spearheads the AI and machine learning initiatives there. She has more than 20 years in the tech industry and has been part of technology initiatives at McDonald's, Canada Life Assurance and Citibank, just to name a few. Viva, thank you for joining us.
Speaker 3 00:12:34 Thanks Kesa. So nice being here.
Speaker 1 00:12:37 Great. First of all, tell us about the role that AI plays in market data, specifically market data analysis that drives decision making.
Speaker 3 00:12:47 Uh, well that is, uh, hidden figures and there's hidden data between the numbers. Um, numbers per se just give facts, but if you, uh, read numbers in association with other numbers, it gives you a flow of information. There is a pattern that comes out through it. And this pattern recognition is market is artificial intelligence, and that's what machine learning is all about.
Speaker 1 00:13:16 So in terms of who uses this, so who, the end user, would they be primarily researchers or would they be primarily portfolio managers who want to understand patterns? Who would be the primary users of this type of market data infused with ai?
Speaker 3 00:13:32 Uh, well, there are two answers to this. I'll, the broader answer is that, uh, it can be used now that, uh, in the digital age when in the digital age of something like Robinhood, uh, wherein everything is so global and everything is so democratized that anybody can do, uh, anything and market, uh, the information that you have is, is instantaneous to everybody. There's no lagging information. Uh, so this information can be used by almost anybody who wants it in the sense that supposed, uh, there is a portfolio managers who want to use it to, uh, to, to buy a certain data. And now you've got these meme stocks coming in like GameStop, uh, wherein, um, it is even given to users of Robinhood. So this can be used by anybody, and I'm talking only about financial data. It can be used by anybody who's interested in buying and selling of stocks.
Speaker 3 00:14:24 So this is in the larger picture. In the smaller picture, uh, the what as with respect to what analytics insight does is what we are doing is we are trying to cover, uh, stocks, uh, across all 49 countries in this world. If you see, most of the stocks that are covered is usually comes out of the US or Canada, and there's huge analyst coverage, but there are also stocks in these, in these countries which are not covered because it's a small cap stock or a, a pre-revenue stock. And there are also stocks all over the world. Like if you want to know some analysis on a stock covered in Romania, well they won't have 1500 analysts covering these stocks because there are only 70 stocks. Uh, what this is where we come in and we do machine learning and ai, and we give analysis, uh, across all stocks all over the world.
Speaker 1 00:15:14 So this is great, and I'm so glad you brought up Robinhood and GameStop. We've talked a lot about the democratization of data and how it is becoming increasingly accessible to more and more people. Are there some challenges or some cautionary tales that we need to take into consideration in terms of ensuring that we continue to, to give data, but also in tandem to educate users of data? Could you talk to us about that?
Speaker 3 00:15:42 Uh, yes. This is very, very important in democratization of data because the first thing that you need to verify is the source of the data. Now, anybody can come as an analyst and anybody can come in as, um, uh, as a person and give their opinion on it. What happened a few year, a few months back was that, um, this herds as a company declared bankruptcy, but it stock went on going up. So in these cases, uh, when, when data is so democratized, there are many people who are not educated enough in, in, in the, uh, act of buying and selling a stock. So in these cases, what happens is if you are the last person in the line, then you are held, you are left holding the bag. So what is most important is you need to verify the source of the data, the veracity and the validity of the data. That's more important.
Speaker 1 00:16:35 And as we move forward, Viva, in terms of looking ahead, understood that right now, understanding the source of the data is the utmost importance. Are there trends going on that you can speak to where in the next couple of years, there are a couple of other pieces that we might need to think about in order to educate ourselves? So if there are trends, what are two or three top trends that you're seeing? Not right now perhaps, but we'll definitely have an impact in the years to come.
Speaker 3 00:17:00 Uh, you're talking about the democratization of data in that sense?
Speaker 1 00:17:05 Yes, in that sense.
Speaker 3 00:17:06 Well, what will happen is, um, right now we are in information overload stage, uh, wherein, um, if there's so much information coming in and it's, uh, and we are not able to make a good, uh, correct decision out of it. So I think this is going, we are the churning stage, uh, right now wherein it's just viewed all over. What will happen is out of this will come out, uh, people, democratic people who are like no single-minded individuals who will be non-biased reporters to, so that we won't have the.com bubble that happened in 2000, in the early two thousands. So the trends that'll happen is there'll be certain people whom you would start following, uh, because they give non-biased information. And, and there'll be public disclosure, not just the usual ones that come in saying that this person is no longer affiliated to blah, blah, blah, blah, blah.
Speaker 3 00:17:57 Uh, because of the information overload that we are having, it'll be possible for, uh, small startups, uh, who are not able to get their space under the sun or get their information out, uh, to the market might be easily be able to, uh, be able to market their idea. Like crowdfunding is something which was never heard of, say, 10, 15 years back. You'll have to go to a person in the market and, and, uh, and raise money for your idea. Now, crowdfunding, something is something that exists. So I presume, uh, it's a double X sword, but these are the few trends that I'm seeing.
Speaker 1 00:18:34 And right now we're talking about your work and your experience specifically in the financial space. I know you've also worked for other types of industries, McDonald's, for example. Is there a difference between data usage and how the data's flowing with financial services versus an another type of industry?
Speaker 3 00:18:52 Uh, yes. Financial services is blessed with the fact that there is an information overload. There is pieces of data floating and lots and lots of data floating, uh, in the, in the data world. The only problem is the data needs to be clean. Uh, the problem with financial data is there's lot of data, but lots of unclean stuff. So if you want to handle financial data, you have to really be a person, uh, who can sit down and clean the data. And if you talk to any data analysis or data mining or any, uh, artificial intelligence expert, they would say that 70% of the time, the data, the time is spent in cleaning data. When you compare the information to other sectors, uh, like say McDonald's, they're, the, uh, data is usually, uh, tied to certain data points. Like, you know, McDonald's would want to know what is moving faster in their restaurants and all that stuff. So it comes from a certain point of sale. So here the data is relatively clean, but it's only one single type of data with financial data, I feel with E S G and, uh, you know, these, uh, these, what do you call it, meme trends and everything that's coming up, the entire democratization of data is affecting the financial world more as compared to others.
Speaker 1 00:20:06 And so we're talking about in the financial world, as you mentioned, clean data is important. As you know, Viva, we're really celebrating opportunities for women in technology, particularly in these startup areas, to really look at ideas and really act on those ideas. If there are folks who were interested in that type of area within technology, so looking at data, cleansing data, that sort of thing, what would you say they would need to look at in terms of processes? What are the top two or three things that they need to consider before they think about going out there and really starting a business around this?
Speaker 3 00:20:42 Well, for cleaning data, the first thing you need to, uh, understand, uh, you need to have a, I would respect to educational qualifications. You need to work with math. Math is something that's not supposed to daunt you. And, uh, the other thing is, uh, uh, I would say like, you know, as a woman, as a mother, as a, as a parent, it is something that comes to us in abundance, that is patience, especially when you become a parent, you become, uh, very, very more patient. So I feel patience is something that you need to have when you're looking at big data and, and volumes of data, mainly because, uh, to find a pattern and to find how, how, how much it's re repeats and cleaning it, and again, doing the same thing again and again. For you to find one small, uh, pattern of data might sometimes take you month, two months, three months just to find a pattern because you, you have to perennially clean the data, remove the noise, go back again, clean the data, remove the noise, and, and do it.
Speaker 1 00:21:39 And for some of us, that involves a different way of approaching a problem or a different way of thinking, right? So to see patterns as opposed to looking at it from maybe a different type of perspective, like a linear perspective, what, talk to us about that. What type of, um, thought process, what type of approaches do we need to really focus on in order to see things in that pattern sort of way?
Speaker 3 00:22:02 Um, the, the main thing is data by itself would be a data point by itself. Like, you know, it's like, uh, uh, uh, it, it's like, it's like marrying space, uh, uh, space to, to, to data as, as you go in space. The more we go and the move deeper, we go into space. And the more deeper we go in there, we find out that, um, till about two years till about a year back, black folks were not supposed to be transmitting any information. But right now, after seeing it for so long, by observing it for so long, uh, we find out that black, black coats do transfer information outside transport, data matter outside when they start spewing it. So similarly, what happens in data is, uh, a data point by itself might be right or wrong, but when a data to observe a pattern, you have to see it in conjunction with other data points extraneous to it, uh, within the own, within the same data set. So it's like there's not much of a difference between observing the world through a telescope and looking at data models, uh, and data.
Speaker 1 00:23:09 Love that analogy. That is great. And people talk to us about your background. I mean, you mentioned math and you also mentioned having patience in the industry. Was was math really your path into this, or did you take an alternate path?
Speaker 3 00:23:24 Uh, well, um, um, in my parents are both science graduates, so science was part of the family. Um, I was very much interested in space. Uh, I'm talking about the early, uh, uh, when the, uh, the Colibri shuttle launched. I think that was an astounding moment in history. Um, when the pc, when the first IBM PC got launched wherein when Microsoft got launched wherein you could use Word and all that stuff. So, uh, new inventions, the new things, which really changed, which created a paradigm shift in the world, uh, really excited me. Um, math was something that was, that, that was a, that I had to work on a lot. But yeah, it was not very difficult, but it was not very easy either. Uh, so that usually got me into the pattern of, of, of technology, like the only way I could go. I was really interested in space, so technology was the only thing that would take me towards it. So I was start interested in, uh, in coding and all that stuff. Um, uh, there was a coding class when I was in grade eight, and my mom thought it'd be a good thing to do in summer. She put me into it and I took it like a duck to water, and that's when my entire journey started.
Speaker 1 00:24:36 What was the language?
Speaker 3 00:24:38 Uh, you won't believe. Um, it's something called, uh, basic. It's, it's a very old language, so yeah, <laugh>.
Speaker 1 00:24:45 Well that's great. And from there, what other languages are you, are you looking at now? Python, I know is, is big in FinTech. What, what's really big now in different industries? From financial services as well as in the, um, in the other corporate industry spaces
Speaker 3 00:24:58 Right now in big data, it's Python or R
Speaker 1 00:25:02 Python or,
Speaker 3 00:25:03 Yeah, those are the two main things, uh, that you use because it, it's, it's mainly used because the, uh, uh, when, when I started coding long time back 15, 20 years back, uh, we had to write all the pieces of code by ourselves. So there was lot of duplication of code. Now if you take Python, what would take me at approximately an hour to code just takes me 10 minutes because all these libraries are, are free for use, it's open source. So I just u call that piece of code and I run my code. So I'm more interested in finding out the final answer rather than using all the building blocks that we had to build in those days.
Speaker 1 00:25:38 And that is a great approach to take. Viva. Let me know if we, you mentioned that you started learning coding as a child. If there are adults who are interested in coding, what would you say to them in terms of the best way to go about it? Where should they start? Which language should they start and what kinds of things should they be thinking about now?
Speaker 3 00:25:57 I think there's no age wherein you are too late to learn to code. That's absolutely because if, as as human beings are programmed to be logical, we are a bit emotional, yes. But we are logical creatures. So yeah, there's never a, uh, even when, if, even if you are a hundred and you want to start learning to code, please do. I would say, and there are so many free resources available, uh, that you can find out. There's so many, uh, uh, courses. I think the first thing you should start off is Python. Cause it's really very easy. Um, and coding is right now very easy, mainly given the fact that the, it's open source and people are ready to share their insights, uh, and, and work as a community.
Speaker 1 00:26:39 That's great. So never too old to learn coding, working as a community to learn it is a great way to do it. Viva, thank you so much for your time.
Speaker 3 00:26:48 Oh, thank you, Keisha. Thank you so much. Great talking to you.
Speaker 4 00:26:54 We invite you to subscribe to the Refinitiv Sustainability Perspectives podcast on iTunes, Spotify, or wherever you stream your content. What did you think about the podcast? Leave us a review on iTunes or follow us on LinkedIn and Twitter for updates on our show. You can even check us out on YouTube now. Thank you for joining. See you next time.
