Welcome to episode #167 of the Super Data Science Podcast. Here we go!
The use of data science doesn’t stop with data scientists only. All operations within an industry or organization could highly benefit from different data science techniques. Learn from Nisha Iyer how your industry could be shaped with data-driven decisions plus a lot more in this episode!
About Nisha Iyer
Nisha Iyer is currently the Director of Data Science at Data Society. Before that, she led teams from Booz Allen Hamilton and Discovery Communication to develop their use of data science. She has already helped different industries and business to adopt a data-driven culture.
Incorporating data science techniques into a company’s daily operations (e.g. marketing, human resources, manufacturing, etc.) could drastically improve their performance. Data Society, an EdTech company based in Washington, D.C. has been helping different organizations and companies incrementally and effectively adapt the practice of data science. Nisha’s company is doing it through hands-on training, online courses, boot camps, etc.
In Nisha’s experience, most of the people in the industry or organization she has helped are receptive to learning data science. You don’t have to be experienced in data science to take their training. The aim is to spread data literacy inside every industry so Data Society offers courses at different levels. Everyone inside the industry or organization should be adept at handling data.
Data Society has built training from simple to complex, and from starters to business executives. Nisha also has clients from federal agencies and private companies. She tells that the demands might be very different in terms of learning and the training. But the fact still remains that data-driven decisions could improve their performance.
Nisha has been in the hospitality and communications route before conquering the world of data science. Her career journey towards gaining expertise in data science is something where everyone could learn from. Nisha tells that it wasn’t that easy to make the transition and she learned a lot along the way. The important thing is the passion and the drive to achieve your goals. Her communication skills that he learned before have greatly helped in her growth in data science. She also talked about the tools R, Python and AWS above the many tools she uses.
Nisha says that she is always excited about the creativity she could contribute in the field. And she still aims to increase her contributions in data literacy. At the end of the podcast, she shared some of her ups and downs as a data scientist, what she thinks the future will be for data science, and many more.
So better listen to the latest episode of the Super Data Science Podcast to learn from Nisha!
In this episode you will learn:
- Nisha tells more about Data Society. (04:38)
- The similarities and differences of the data science culture in a government setting and a private company. (07:30)
- All employees should be adept with data science, not just data scientists. (09:08)
- Nisha shares stories on how Data Society drastically helped their clients. (11:28)
- Attending hands-on training taking an online class. (14:12)
- Learn how to reach out to more clients and communicate more about data science. (25:02)
- Your data science career benefitted also from her previous experience in other fields. (25:02)
- Transitioning to a different field (i.e. Data Science) will be challenging but definitely worth it. (34:55)
- Nisha tells on what’s the most impactful tool she used. (38:08)
- It’s important to learn what happens in the algorithms to better communicate the findings. (41:40)
Items mentioned in this podcast:
- Data Society
- An Introduction to Statistical Learning: With Applications in R by Gareth James and Daniela Witten
- The Elements of Statistical Learning by Trevor Hastie and Robert Tibshirani
Kirill Eremenko: This is episode number 167 with director of data science at Data Society, Nisha Iyer.
Welcome to the Super Data Science podcast. My name is Kirill Eremenko, data science coach and lifestyle entrepreneur. And each week, we bring you inspiring and ideas to help you build your successful career in data science. Thanks for being here today and now let's make the complex, simple.
Welcome ladies and gentlemen back to the Super Data Science podcast. Super excited to have you on board and today, we've got a very interesting guest, Nisha Iyer who is the director of data science at the Data Society. So Data Society is a company that is based in Washington and they help businesses adopt a data science culture, a data culture in their organization and enhance the literacy of their employees. They work both with companies and federal agencies. And so what I found interesting about this conversation was just the whole notion that with time, more and more organizations are going to require more and more people in their companies or within their operations to actually be familiar with data science or even more, be able to operate with data.
And the interesting part is that this is not a one-way game. As Nisha shared, you will see that people and organizations want to do that as well. People in organizations want to learn and one of the examples Nisha shared was that there was an organization where they were doing some training and only had 40 places available but something like 270 people applied to be part of that training. So it just stands to show that employees and people who work in companies actually recognize the importance of data science, regardless of their actual role, regardless of what they do. They just see that it is important and it does make a difference, it does add value to their roles. So that's what we're talking about in this podcast, plus you'll get to know Nisha's background, where she came from. She's actually from a background that's not technical. She has a hospitality and communication background and still nevertheless, she got into data science. So a very inspiring story of all the challenges she overcame to get to where she is. And on that note, not going to hold you up any longer. Let's dive straight into it. Without further ado, I bring to you, Nisha Iyer, director of data science at the Data Society.
Welcome ladies and gentlemen to the Super Data Science podcast. Today, I've got a very special guest, Nisha Iyer, director of data science at the Data Society. Nisha, welcome to the show. How are you today?
Nisha Iyer: I'm doing well. Thank you for having me. I'm really excited to be joining you guys.
Kirill Eremenko: I'm very excited to be having you. And tell us where you're calling in from.
Nisha Iyer: I'm calling in from Washington DC.
Kirill Eremenko: Washington DC, amazing. What's the weather like there? You mentioned it's 18 degrees or something, 65 Fahrenheit.
Nisha Iyer: Yeah. Yeah, it's cool today but it's usually muggy and hot in the summer. It's summer here right now so we got a nice break from it.
Kirill Eremenko: Nice, nice. Well, it's always good to have a bit of variety, isn't it? In the weather. Not too much though.
Nisha Iyer: Yeah. Well, we have a lot of variety. DC gets all four seasons definitely but the summers here get pretty hot.
Kirill Eremenko: Okay, that's good. What I was referring to more was ... We have a city in Australia called Melbourne. If you go there, you get four seasons every day.
Nisha Iyer: Oh, yeah.
Kirill Eremenko: Yeah. So people listening from Melbourne, I know you hate me for it, these comments. But every time I go there, it happens. You don't take an umbrella, it rains. You take an umbrella, it's hot. It's very interesting climate there.
Nisha Iyer: Unpredictable.
Kirill Eremenko: Yup, exactly, exactly. All right. So Nisha, you are the director of data science at Data Society. It's an EdTech company. Tell us a little bit more about Data Society to kick us off please.
Nisha Iyer: Sure. So Data Society is EdTech startup. And what we mean when we say EdTech is education about data science. So our goal, our main goal is to spread data science literacy and to help companies spread data science literacy throughout their companies. And as we're situated in the DC area, we work with a lot of federal government agencies as well as private firms. So what we do is we work with these companies to build customized trainings for them so that they are able to train their employees to be able to use data science better and actually make data-driven decisions.
So I think that this came from just the observations from my manager and co-founder of the company. The two of them just saw that there was a lack of actual data science being used in companies. It's such a new, emerging field, especially in this area and everyone ... It's like the buzzword. The buzzword's big data, predictive analytics, machine learning and then, at the end of the day, no one really knows what to do or how to use it and it just becomes the words versus actually looking and using the data to drive the decisions. So what we do is we build trainings as simple as introduction to R and Python and as complex as deep learning, neural networks, and we deliver those trainings with our in-house instructors, and our in-house instructors go and deliver trainings to all the different places we work with.
Kirill Eremenko: Okay. Wow. That's a really cool company. A very cool mission. How long has the company been around for?
Nisha Iyer: So the co-founders started it in 2014 and I think it's ... I joined last year, so I've been with the company for about a year. And when I joined, there was about four people and now we have ... There's four people working on the actual Data Society. We work out of a coworking space so there's four people here. There's about 20 people total because we had some people remote for the development side and stuff, and now we have about 12 people sitting in our offices, so it's growing fast. It's pretty cool. It's been pretty cool to see the company develop.
Kirill Eremenko: Wow. So it's grown threefold in just a year. That's insane.
Nisha Iyer: Yeah.
Kirill Eremenko: Yeah, that's very cool. Okay. And so just stands to show how valuable this kind of mission and approach is. And you mentioned you worked with both companies and government institutions being in Washington DC. What would you say is ... Are there any major differences in terms of demands for your product or in terms of what it is that you help them with mostly?
Nisha Iyer: I think the similarities are what we help them with mostly because all of them just need help in ... Many times, it's different subject matter or it's a different approach to the type of training. Like for example, we're currently working with health and human services, which is the federal agency, HSS in DC, and they want an eight-week bootcamp delivered. So it's spread out over eight weeks. Many of our corporate clients, we will do one-week long bootcamp. So it's a rigorous type of course that every day is eight hours and we deliver data science 101 in one week and they go through a lot of different ... And then they take that knowledge and apply it. Whereas the federal agency, we work with them to build a capstone project over the eight weeks. So maybe it's just in the difference of delivering but overall, I think that everyone is looking for the same result which is how to employ data science better within their agency or within their company.
Kirill Eremenko: Very, very interesting. And so is my understanding correct that you don't just train up the data scientist per se, you train up pretty much anybody or even everybody in the company to be more data literate and to be more adept with handling data or leveraging the power of data in their day-to-day roles?
Nisha Iyer: Yeah. There would be a technical ... For example, the training we're giving currently for HHS, we had an interview process for people to get screened to come into the course. So we actually had a huge turnout of people that applied for the bootcamp. About 475 people and the cohort can only be 30 people, so we had to whittle down the process by seeing who would best fit in. So with that, we actually looked for who had a bit of a more technical background so they'd be able to catch on faster. So these people, this training is directed more to people with some type of technical background or maybe they've done some kind of programming in their past but they don't have a very strong foundation and they also don't really have a foundation in data science, They don't know the algorithms and the statistics. They're missing one of the three. The intersection of the three areas that kind of make up data science which is, I guess, the business application, programming and statistics, and that intersection of that Venn diagram is what we think as the data scientist.
So yeah, I think that ... We also do have trainings that are geared for executives or for people that want a higher level of understanding of data science and they can just understand the jargon better or talk to the people they manage and understand what they're actually telling them. So yeah, we have all different levels of training.
Kirill Eremenko: Got you. And you guys focus predominantly in the Washington area or is this national?
Nisha Iyer: It's national. I mean it's international. We've had clients contact us internationally and we are open to whatever comes our way right now. Just as a small company, we are ready to take on the world.
Kirill Eremenko: Nice, nice. Great attitude. All right. Okay. And so what has been the feedback so far from companies that you've done this bootcamp for or these executives or other types of trainings? Can you share some of the success stories that your clients have had?
Nisha Iyer: Yeah, sure. There have been great success so far. Just in the time period that I've been here, I've seen trainings be deployed and then the clients coming back with the information that's been taken from the trainings being used in their day-to-day work, and there's multiple stories like that. One that really resonated with me is, again, the HHS bootcamp that happened last fall were the first pilot of the bootcamp, the eight-week bootcamp I was telling you about. So one of the students or one of the people in the class, he worked for the NIH, the National Institute of Health. It's under HHS. And he worked in the department where they had to look through documents, look through many documents to find ... It's kind of like a discovery-type process, looking through health documents, looking for certain statements and it was a very ... Just a very drawn out process.
And so what he said, he presented the problem statement as this person takes 13 hours to go through all these documents and he wants to make it a shorter process. And so what we learned in the class, one of the big part chunks of the eight-week training was text mining, text processing. And through learning that and through building in some of the predictive algorithms, the predictive methods we've taught, using both the text mining and predictive methods to classify documents, he was able to take that 13-hour process down to under five minutes. So instead of taking this much time to do this process, they're going to be able to do it that much faster and get through so many more documents every day. So that was something that was just so cool that we can ... And that's also reiterate the point that these people are ... His position in the company was something closely related to data science or data analyst but he still didn't have the tools to make him "dangerous". He wasn't able to do what he was able to do until we went in and kind of ... Where actually you could be doing so much more with the data and the tools that you have at your disposal, so it was pretty cool.
Kirill Eremenko: Interesting. Very interesting. I'm loving where this is all going. That was a very cool case study. But I'm going to play the devil's advocate here, I'm going to ask you a question. Because as you probably know, Super Data Science and myself, we're in the space of online education. So my question to you will be how is your training more beneficial than somebody going and learning these skills online on their own? So what advantages or what additional value do you bring into organizations? Why would an organization hire you guys rather than just asking their people to, "Hey, guys. Go and learn data science online"?
Nisha Iyer: Well, honestly, I think the online trainings have a great ... I've taken online trainings myself. I did my masters in data science a couple of years ago and as a student, while doing my masters, I knew exactly where to go, exactly what I needed to learn. So going online, I was like, "Okay. I need to understand the basics of progression better." Or, "I need to understand this part of clustering." Or, "I need to understand ..." So I would go google that, find a course and then take it and understand what I need to understand. But what we think that we offer that is different because we're in the classroom and actually interacting with the students is that for most of these people that we're approaching have little to no background in actual data science. And what we've heard from the companies that we've worked with or the federal agencies is that sometimes online, they can just get lost. There are just so many different platforms, there's so many different courses that there's not a straight trajectory where they can be like, "Okay. We are going to learn A, B, C, D and then we're going to have what we need to move forward and be armed with what we need to continue moving forward."
So I think that this isn't, "Okay. This is better than online training." It's more like this is a compliment to online training and this is something instead of going to a two-years master's degree program, which many of these people might not have time for, why don't we just get your feet wet and have a quick deep dive into data science and give you the tools to understand what you will need to move forward with. So let's say this guy, NIH, he knows what he can do to build out this text mining tool but now he also knows what else he could learn in data science and now he might go online and look for more resources. So I think it's just the hands-on training is something that is ... For me, I've taken both online and I've been in classrooms because I was doing my masters and I think that sometimes, being in front of an instructor is also very helpful.
The way we put our trainings together, we try to keep a couple of case studies and take them throughout the week or throughout the eight-weeks so that students see the way different algorithms react with the data and we also just make sure that we're sometimes customizing it towards that subject matter. So for HHS and for NIH, we have a lot of public health type cases and for another company that we work with, they're a huge defense contractor, so we'll have more to do with engineering and science. So we're able to kind of get the people's attention because that's what they're working with. So I think those are some of the major differences.
Kirill Eremenko: Got you, got you. So you have the tailored approach to specific businesses or industry's demands. And on the other hand, you're not comparing, contrasting this to online education, you're actually complimenting online education by kickstarting their process and giving them specific steps they can follow and giving them a feel, a taste for what power they will have after they ... Or the power they will gain on this journey into data science. Okay. Makes total sense. Thank you for that answer. Okay. So that's what you guys have achieved, accomplished since 2014 and this massive growth in the past year is very impressive. What are the plans for the company going forward? Are you guys expanding into other niches or trying to maybe establish another office somewhere? Is there anything, of course, that you can disclose that is not trade secrets? Any plans for the future?
Nisha Iyer: I think they're all top secret. No, I'm just kidding. Yeah. I think that we're just ... We're really excited about the direction that we're going and the ability to work with some of these large companies. And I think in the future, we ... So something that we've actually employed, I lead a team of data scientist and we actually build out the content for these courses so that's my main focus. It's to organize and to ... Right now at least because we're just starting out and we're really ramping up with all these new clients, so it's to streamline the way that we build our content and that's something exciting that we've kind of dove into and that we have a technical process where we're building all of these stuff in a Markdown interface. We're able to scrap out the code and that kind of thing, so we're hoping to eventually go into ... And this is the maybe more top secret part, but eventually go into also building our products. But we're definitely focused on ... The area that we're focused on is to make it easier for people to work with data science and once again, to spread data science literacy. If that means to building products to help trainings being deployed, maybe that would be something that we're going to be working on in the future.
We also do have a small consulting branch, so we do work with companies to help them on their data science problems. And so as we grow, we can also hope to grow that area and build out our consulting services and see where ... I think it's kind of just like the world's our oyster. We are just so excited to see where this leads and the data science space is growing so fast too. I think an area that's really exciting for me is deep learning and as that space is growing so much, hopefully ... We're thinking we might be starting to build out some trainings there and see where that takes us.
Kirill Eremenko: Got you. No, makes sense. Great perspectives. And one more question I had on this is what's your secret to getting clients? Because there's probably lots of people listening who are like, "I want to be a consultant in data science and I want to help companies in my region or in my country." But they find it challenging to approach companies or approach enterprises on helping them build out this [inaudible 00:21:21]. Are there any tips you can share with people again?
Nisha Iyer: Well, I don't necessarily head up the sales process but I've seen my manager, the co-founders and a couple of other people on our team that do great work. And I think that our secret is just to show our ... We usually will go do a couple of demo trainings and people either latch on or get really excited about what we have to offer, and we just kind of demonstrate the difference that we can make. So we send out information or PowerPoints or decks about our company. I guess that's usual marketing tactics but a big thing that I've seen that Dmitri does is goes and actually just demos trainings, and so that's really helpful.
Kirill Eremenko: Yeah, okay. Cool. That's some good advice. Of course, sending out decks about the company but also some demo trainings, it's very powerful-
Nisha Iyer: And we also ...
Kirill Eremenko: Yeah, sorry.
Nisha Iyer: Sorry?
Kirill Eremenko: You go.
Nisha Iyer: I just wanted to do a quick shout-out. We also have a ... We try to keep a social media presence and we also have multiple blogs, blog posts and our most recent one ... I think I can send this to you, right?
Kirill Eremenko: Yeah, sure.
Nisha Iyer: Hold on. Let me see. So we have a survey. I've just sent that to you.
Kirill Eremenko: Yup, I got it.
Nisha Iyer: That is a blog post about data science communicator toolkit survey. So if anyone of the listeners actually take that survey where it'll direct you to what we put together as the data science communicator toolkit and that is to help with just what we've been talking, what I've been talking about, help with communicating data science literacy within your company. So something that made me very excited about this job when I first heard about it was that I worked at Discovery ... I think Discovery ... Are you familiar with the Discovery Channel?
Kirill Eremenko: Yeah. [inaudible 00:23:34].
Nisha Iyer: Yeah, I worked at corporate in Maryland at Discovery and I was on the data science team. I was managing a data science team and you would think that an organization like that, it's international TV and media organization, would have some kind of data science presence and I was amazed at how hard it was to communicate what data science actually was to the upper-level management and so this toolkit is exactly for that. If you are a data scientist and you're having a hard time working or trying to help people understand why this is so important or that what you're actually doing isn't data science and you need to kind of change your model, then this would help with that.
Kirill Eremenko: Okay. Well, thank you very much. I have the link here. We'll include it in the show notes, so if anybody wants the data science toolkit, that's where you can get it.
Nisha Iyer: Awesome.
Kirill Eremenko: Okay, great. So that's in a nutshell, Data Society. If anybody is interested, anybody especially in the Washington area or national, international, it's DataSociety.com. You can find out more details there. Let's move on a little bit or should I say move back a little bit and talk about your career. So you've had a very interesting journey through lots of different positions, mostly recently to do with data science, and management, and consulting, and so on. I would love to hear about where it all started and how you went from there. Do you mind sharing with us?
Nisha Iyer: Yeah, sure. I definitely have an interesting path. So I'll start way back. I was in the restaurant industry for a while and I was ... I really loved the service industry. I was even thinking about going into culinary and that ended up not working out. But I always think ... I don't know what ... The restaurant industry is different in every country I guess. Because I know when I go to India, it's just ... Here, the servers work on tips so it's a very service-based, customer-first type mentality and I thought I learnt a lot from working in the service industry. I learned a lot about dealing with people and about how to read people, and it's a really great experience. So with that experience, I really thought I wanted to do something with communications. So that's where I kind of went into my bachelors and that's what I studied in my undergrad was communications, which is not a quantitative field.
There was a couple of quantitative classes we had to take in communications and I realized that I really loved statistics. There was a couple of classes about building true experiments, that kind of thing, and I thought I should be focused on something more statistical and more quantitative, and that's where I ... I was looking into masters programs and this is around 2013 time, so big data, all that was just really coming up. And I found an awesome program at the George Washington University that was a masters in business analytics, so it's a combination of employing data science within business and I started that. I was also working at Booz Allen, which is a consulting firm here, and so I got to kind of ... I was doing my masters and working full-time, which is a bit of a crazy experience, but it was awesome because I got to actually employ the things that I was learning in my masters on my day-to-day job and that helped me learn even faster. And then, I just kind of got-
Kirill Eremenko: Just talking about ... Sorry to interrupt. Can you tell us a bit about Booz Allen Hamilton? What does that firm do?
Nisha Iyer: Yeah, sure. So it's a large consulting firm. There's many locations all around the country and what Booz Allen did was just work with, once again, the private sector but mostly, government contracting. I worked on multiple projects with different government agencies and you just go in and you kind of help ... Booz Allen will win contracts and help out the government. So sometimes, there might be a data science ... So we have subject matter experts that go in and help with areas such as ... I started off as the strategic communications consultant, so there's a lot of things like change management and other types of communication. And then I moved into the data science sector which was ... I worked at the FDA actually when I was doing my data science.
Kirill Eremenko: Okay, okay. Got you, got you. All right. Thank you. Thank you very much. And where'd you go from there?
Nisha Iyer: And then after Booz Allen, I had graduated from the masters program and then I went to Discovery, which is Discovery Network. There's about 13 different networks. It's international and it's really cool. It's a really cool company. It was a really cool company to work for and I went into data science there. I started off on the television side and I then moved to the digital side. So I managed the data science team and we worked on optimizing the consumption of the Discovery Network digitally. So we had a bunch of data about all the different viewers that were watching Discovery and we tried to optimize and ... I worked with the marketing teams and the ad sales team to help spread Discovery throughout the world.
Kirill Eremenko: Awesome. Any major breakthroughs while in Discover?
Nisha Iyer: Honestly, my major breakthrough at Discovery was realizing how little people knew about data science, so the breakthrough helped me come to Data Society.
Kirill Eremenko: Yeah, okay. Got you. All right. So now you're in the business of fixing that issue.
Nisha Iyer: Yeah or trying to help companies, yes.
Kirill Eremenko: Okay, interesting. And so what would you say have been ... Apart from that one in Discovery, what other major learnings have you had on this pathway? Because, first of all, it's very inspiring to see that you came from a very unrelated, I would say, background in hospitality and then communications, and you made your way to data science. It's got to be very inspiring for our listeners who are coming from other backgrounds as well. But what would you highlight as your major learnings along the way or major ... I don't know. Career, again, breakthroughs that have shaped the professional that you are today?
Nisha Iyer: Yeah. That is a big part of my story that I think is so important and I hope that people that are listening that haven't started off in their careers as a technical background is that there's no wrong time to start doing this. You just have to have passion and the drive and that's what I had. That's where I really found that this is something I loved. I love data science. I think that it can be ... My love for it is that I think that there is so much out there that hasn't been done and that's kind of where my excitement comes from like, "Oh. There're so many areas. There's so many subject matters or different sections of work that don't use all the data that they have." And I think that my career breakthrough was finding that this was my passion and then I went full force driving ... And that's what I was always looking for, what was I really interested in. And when I found something new like for example when I was studying communications and I realized that maybe I actually liked this quantitative part more than the written communication, instead of backing down because I didn't have enough of a quantitative background from my university studies, I just went forward full force.
And let me tell you. When I started my masters, it wasn't a walk in the park. Most of the people I started my masters with had computer science of mathematics undergrad, graduate degrees and I was there with a totally non-related, like you said, field. And I walked into my first class and it was Python and my teacher just started going off about for loops and functions. I had no idea what he was ... Yeah. I was just like, "Oh my God. What have I walked into?" And I just went home, I used online learning to learn Python. I went on and just had to spend extra hours. And I'm struggling a lot more than some of those people but at the end of the day, I walked out feeling like I learned so much and that just changed my career. The things I learned in my masters and kind of flipping around into being able to use the ... I think what really helped me was being able to use my communication skills which I've learnt throughout, through the hospitality industry and then into my undergrad. Use those communication skills as well as learning the hard technical skills and being able to kind of merge the two.
Kirill Eremenko: No, [inaudible 00:33:19].
Nisha Iyer: Because what I ... Yeah, because I felt what I saw when I was doing my masters and bumping into all these computer scientists, and mathematicians, and people that knew Python and knew all these statistical theories and stuff, but they couldn't ... They were having a hard time being able to communicate it to let's say someone that has no idea what a for loop is or has never programmed before, and those are the people you need to communicate these stuff too because those are your upper-level management most of the time.
Kirill Eremenko: Yup. Yeah. It's very interesting entry into the data science field when you come from a non-technical background. Like in this case, leveraging your communication skills. Because if you do have the perseverance and determination to power through and pick up those communication skills ... Not communications. Those technical skills, those programming skills and other things, and algorithms, and all the maths and statistics. If you do manage to do that, then all of a sudden, you find yourself with this superpower that you actually came into data science with at the start which is communication skills and that really puts you ahead of the game. Because as you said, for instance that's one of the most important and also lacking areas of expertise in the field where people can actually communicate these insights. So in this case, your career is definitely a testament to that. Anybody can do it. But my question will be what is the driver for your determination? I understand the passion and you noticed that you love data but then walking in to that classroom, sitting there and being bombarded with all these for loops and Python and so on, where did you get the courage and the strength to actually persevere and continue learning these things that must have felt extremely complex at the time?
Nisha Iyer: Well, that's a more personal journey but I guess that I can just say I've had a lot of ... Throughout life there's been a lot of ups and downs and before I got into this career path and before I even got to pursue this masters, it was something I never thought I would do. And so being to even sit in the classroom and have the opportunity to learn these things, I didn't want to just go at it with half measures. I wanted to push forward and just do my best. And maybe it's competitiveness. I don't know. I know personally I'm competitive. I play a lot of sports, I get really competitive and it seemed like the same kind of thing in the classroom. I'm not going to not do this. I have to do whatever it takes to succeed and then that's where also the passion also kicked in because it was like an interest like, "Oh my God." I'm sure if you're listening, you've also experienced this where you learn about a certain algorithm and then you start reading about it and then you start getting all these new questions about what does this mean, what does this mean, and that's how I spent so many nights. Just waking up ...
It's so great that there are all these online sources for learning about data science because I would just spend hours, sometimes just in rabbit holes of learning more and more about certain things, and then ideas would come up in my head like, "Oh. I can actually use this to work in my current job. It would help build things. It would help optimize production." That's where the courage aspect would come in because I would have to take these ideas that could be total flukes and I would talk to my manager about them. I think I realized that a lot of people just don't bring their ideas up and if I do, worst case scenario, people are going to shut me down. Best case, I get to be heard and I get to improve something within the company and that's kind of what I pushed forward on since then.
Kirill Eremenko: Okay. Well, thank you for sharing that. It's very inspiring. I'm sure lots of people will be able to relate to that. And as you said, we've all had those times in our careers when, like you, learn something and you have more questions and more questions and ... Sometimes, I personally just get so deep into these questions all the time and I look up from my computer at some point, it's midnight and I'm like whoa.
Nisha Iyer: Exactly.
Kirill Eremenko: Where did the day go?
Nisha Iyer: Yeah, exactly.
Kirill Eremenko: Yeah. Okay, all right. Well, that's a very inspiring and exciting journey into the field of data science. Let's talk a bit more about the technical side of things. So going from not knowing anything like R programming to where you are now. I'm just going to read out your data science toolkit that you have on LinkedIn. So guys, get ready for this, "R, Python, Spark, Hadoop, AWS, Pig, SAS, Bash, Tableau, SQL, NoSQL, R Shiny, Data Mining ..." And more. How insane is that? That's just unfathomable for me to ... I wouldn't be able to pick up all those tools if I came into the field not knowing anything. So tell us a bit more about what tools ... I know that probably now in the educator space, you don't use them for actual consulting projects. Maybe you do, maybe I'm wrong. But throughout your career, what has been your favorite tool? What has been the most impactful tool that you've used? Any tools that stand out, that pop to mind?
Nisha Iyer: Throughout my career, I guess it depends on where I was working and what I was doing. But I think what my favorite ... My bias would go towards the things that I think would be the most helpful to learn is R and Python and I think there's always an ever going argument about, "Oh. Which one's better?" I don't think either one's better. I think they're both better for different things.
And also at Discovery, I had to use AWS a lot, Amazon Web Services, and I think that was amazing to work with large amounts of data. The processing power and speed at which we're able to do things is ... It was definitely greater than what I'd ever be able to do on my local machine, so that's something else I really enjoyed working with on consulting projects. And the ease of setting up the virtual machines and the ease of setting up clusters and being able to use cloud computing, it was great. It was something that I really thought was efficient and also just working ... We used Redshift which was where we stored most of our data. And just the way that we're able to pull in data from web sources into the database was great as well. We're able to set up processes so new data came in every night and it's something that I really ... Outside of the teaching, education space and just day-to-day data science, something that I really think that helped me in my career.
Kirill Eremenko: Okay. Got you. Interesting. So R, Python, AWS. Okay. And so are these tools all the tools that you guys teach at Data Society?
Nisha Iyer: Yeah. Most of our courses on focused on actual data science execution and building algorithms and stuff, so we're usually building out our trainings in Spark and Python. Sorry. R and Python, but we do ... Yeah. If a company says, "Oh. Actually, can you guys also help teach how to deploy Spark? Can you teach how to use R Shiny? Can you teach us how to connect SQL with Python or can you ..." We have an intro to SQL course. So like I said before, we kind of customize our trainings. So our major platforms are R and Python and then we kind of build on those to whatever our clients are asking for.
Kirill Eremenko: Got you. And you mentioned that the two probably most important tools to learn are R and Python. What do you have to say for those listening who really don't enjoy programming or don't want to get into it? I know it's not a common thing I guess in data science but sometimes, there are people who appreciate the power of data science, appreciate the power of speaking of data and so on but don't really want to go down the pathway of machine learning or of R and Python. Have you encountered those situations in your trainings and what is your usual approach in those case?
Nisha Iyer: Yeah. I think if you don't enjoy the programming aspect, I think that a place that you could kind of see yourself within data science would be more on the business side and being able to understand what's going on but not having to actually deep dive into writing the code out. And so I think some of the better tools in that case would be something like Tableau or building out the visualizations. So Tableau is pretty user-friendly. There's a lot of things that you can do with the graphical user interface. I do think that a lot of the cooler Tableau dashboards that you have to put together, you actually have to use some hack type of [actions 00:43:10] but those are some tools that you can use to actually work with data that's already been manipulated, aggregated and built out. And I think that if you don't actual enjoy the hardcore, the programming part, it's still really useful to understand what is going on within the algorithms so that you can communicate the findings to people that have no idea what data science is about.
Kirill Eremenko: Okay, got you. Thank you. I totally agree with that. So visualization tools or understanding it from a business perspective, communicating data sciences. There's lots of space for any kind of approaches you want to take to data science. Okay. How about a rapid fire list of questions? You up for that?
Nisha Iyer: Yeah, I'm ready.
Kirill Eremenko: All right. So what is the biggest challenge you've ever had as a data scientist?
Nisha Iyer: That's a hard one. I'd have to go back through my list of all my projects that I've worked on.
Kirill Eremenko: What's the first one that pops to mind?
Nisha Iyer: Honestly, the first one that pops to mind is something that I'm working on currently and that is just being able to ... It's not as much a data science problem. Not as much a data science challenge as ... So my day-to-day role today is directing a team, so my biggest challenge is optimizing production of how fast we build these trainings and being able to build ... So we build out eight-hour trainings, four-hour and eight-hour trainings. And my biggest challenge is being able to fit in the perfect amount of information that's technical enough, but not too technical to overwhelm students, but also not too vague because we want to really get into the weeds of data science and that's a huge challenge. Because when you're trying to explain something like principal component analysis, how do you stay non-technical enough without getting into exactly what's going on with PCA and under the hood?
So we actually work closely with the team of instructional designers that we have at Data Society. So my challenge is that the data scientist I work with and myself sometimes want to get very technical and the instructor designer team wants to stay more on the broad side because we can't fit all these material, so just the merging of both of those sides. On the whole, that's kind of the data science challenge. When you're taking these data science projects to stakeholders, you want to stay out of the weeds of the technical part but also be able to explain the impact that there is on the business. So that's how it related outside of actually building the trainings but yeah, my biggest challenge right now is that exactly. It's how do we get this technical enough but keep it simple enough so people can digest it.
Kirill Eremenko: Okay. Got you. Great challenge. I face that challenge on a monthly if not weekly basis as well. Totally, totally I can relate to that. All right. Next one. What is a recent win you can share with us that you had in your role? In your current role, something that you are proud of.
Nisha Iyer: I think the recent win, I did discuss earlier but it's just that ... So when I came in, a lot of our material was actually not in ... It wasn't done in Markdown. It was a lot of decks that were together where we had code put into the decks with copy and paste from the R script into this text box and there's a lot of formatting and that kind of thing. And I think the biggest win that I've had and that I'm really proud of is that our team has come up with a way that we can write everything, the code, the content in one script and then we built a scraper-type thing that actually splits that script and it builds the slides out in a Markdown format using Reveal.js and then it also builds out the code, the exercises, and so it's a streamlined process. So if the code changes which code always changes because there's always updates and new packages and everything, we go into one place and we don't have to change it anywhere else. And then we re-scrap the module and it'll be ready. So this makes it very easy for us to continuously update with the newest packages, the newest technology instead of having to go and in and change 10 different things sources.
Kirill Eremenko: Got you. So it sounds like you're applying data science to education about data science.
Nisha Iyer: Yeah, pretty much, which is important. Right?
Kirill Eremenko: What's it called? Talk the talk, walk the walk. Or something like that.
Nisha Iyer: Yeah, we have to be doing that. I mean that was my whole point when I got in here too. I was like, "We have to be practicing what we preach."
Kirill Eremenko: Yeah. Yeah, okay. That's a pretty cool ... Okay. I can see why that's a big win. All right. I think you mentioned this a bit but nevertheless, let's maybe expand on it or have another one. What is your one most favorite thing about being a data scientist?
Nisha Iyer: My one most favorite thing.
Kirill Eremenko: Mm-hmm (affirmative).
Nisha Iyer: I think I have a lot of favorite things.
Kirill Eremenko: Oh, okay. Well, what's the most favorite out of all of that?
Nisha Iyer: But I think my most favorite out of it is just the excitement of how new the field is and how many areas it hasn't been applied to and the creativity that comes with data science.
Kirill Eremenko: Totally. Totally agree with you. What would you say is the field that you're most excited that data science is underapplied in that field and you can't wait for yourself or for others to get in there and start applying data science?
Nisha Iyer: Personally, I think that hospitality industry has not enough data science applied to it and I have some personal ideas that I think could be applied to just increase the amount of sales that restaurants have every day, so that's an area I'm really excited about. And there's two. So that would be one and then another one is education. My mom is a teacher and I've heard a lot. She helps with academic therapy so she teaches students with dyslexia and there are some opportunities in that area, especially involving [inaudible 00:50:19] direct admission and diagnosis of people with learning disability. I think data science could help so much with and I'm really excited to explore those areas.
Kirill Eremenko: Got you. Okay, sounds good. A bit about the hospitality, you said you had a couple of ideas. Are those some that you don't mind sharing with the podcast or are those, again, top secret ideas you want to work on one day yourself?
Nisha Iyer: Yeah. Those are kind of in stealth mode ideas but hopefully, one day, every one will be hearing about them.
Kirill Eremenko: Okay, okay. That's really cool. Definitely. All right. And where do we get to? All right. So from ... And I know you mentioned in your bio that this is an interesting topic for you, so here we go. From where you are, from everything you've seen in the space of data and data science, where do you think the field of data science is going and what should our listeners prepare for in order to be ready for the future that's coming 3, 5, 10 years from now?
Nisha Iyer: I think that data science is ... I think that it is a field ... One of my friends was actually saying this. She works at the state department and she had been told that ... She interviewed for a new position as an analyst for foreign affairs and it's traditionally a position where a lot of reading is done, a lot of research and she was told by the people that interviewed her that she needs to have data science skills. She needs to be able to analyze data and so I think that's where it's going. I think it's going to penetrate every industry. I think people are going to start realizing how important the skillset is and it's going to be something ... She was saying you're going to need to know data science. People used to need to know how to use Microsoft Word, something that simple. That's what data science is going to be in a few years.
If you don't understand how to read or analyze data, you're going to be missing out on a lot of information that's available to your company. And so I think that data science, I think that it's just a great area to be in. I think it's a great time to pick it up. It's something that is able ... There's resources out there for you to learn. And anytime somebody asks me about data science, I just encourage them to get online and to just try out a course. And to see if they like it, to see what aspects they like about it. To read a book, to get involved in some type of ... In our area, in the DC area, there's a lot of meetups that are related to data science. So just to immerse yourself into data science. I think that it's unknown where the industry is going but I think that it will be spreading all over, across industries. So it's not just going to be its own subject matter, it's going to be something that all different subject matters use to help them perform better.
Kirill Eremenko: Okay, fantastic. Thank you. That's a great answer. For some reason, I kind of feel like we're Mythbusters right now. There's this myth that data science is just a hype and it's going to drop of after a certain time. Would you agree that we've just debunked the myth and that it's going to continue, data science is here to stay?
Nisha Iyer: Yes. I 100% agree and I think the big debunking of the myth is just to understand what data science is and not to think that it's just some type of mystery or magic or some random prediction that is hype and it's going to leave. There's so many different aspects of data science. I think that's what people need to understand.
Kirill Eremenko: Okay, got you. Definitely, definitely agree. Okay. Well, this nicely brings us to the concluding part of our podcast. I guess one other thing I wanted to ask you is you've already learned so many different tools, so many different techniques I'm assuming and ways of application of data science, and I'm sure that your education is not going to stop there. Is there anything else that you are personally forward to learning in the space of data science?
Nisha Iyer: Yeah, definitely. I think that I'm really excited about learning more about deep learning. There's just so many different types of models and I don't ... I've just only touched TensorFlow and the other packers that sit on top of TensorFlow, which is just one of the many frameworks for deep learning. So I think there's so much out there with that and the different types of image recognition, text mining, all of the different areas of deep learning. I've been reading about it and there's just so much more out there, so I'm really excited about that.
Kirill Eremenko: Totally, totally agree. Okay. Well, on that note, thank you so much for coming and sharing your insights on the show. If our listeners would like to get in touch with you and learn more about your career and where it's going and some of the amazing work that you're doing, what are some of the best ways to do that?
Nisha Iyer: Yeah. Everyone could reach me at my email which is [email protected] And I don't know if you can spell that out but it's just my first name, N-I-S-H-A and then my company, DataSociety.com.
Kirill Eremenko: Got you. And what about LinkedIn? Is that a good place to get in touch with you?
Nisha Iyer: Yeah, sure.
Kirill Eremenko: Awesome.
Nisha Iyer: LinkedIn, you can connect with me and message with me and I'll be happy to be in touch.
Kirill Eremenko: Awesome, okay. And I have one more question with you today. What's a book that you could recommend to our listeners to enhance their careers?
Nisha Iyer: So I was thinking about this because I knew I was going to have to tell you what book, but my favorite book ... And there are so many books out there but a book that really, really helped me at the beginning of my career with data science, I think it really goes in, gives you a good foundation about the basics of data science and it actually is written with code in R, is the Introduction to Statistical Learning. I think that it covers a bunch of machine learning algorithms. It gives you such a good basis and you can ... I find myself referring back to it even today. So I think that I would highly recommend that. And then it has a second book that goes with that after you've got your basis is The Elements of Statistical Learning. And I was in a data science ... One of my masters class and my professor referred to it as the bible of statistics. So I think both those books are amazing and I think I would suggest the readers to ... For you guys to go check them out at least. And you can find them online and download PDFs or you can even get the hard copy.
Kirill Eremenko: Got you. All right. Thank you very much. Those books come up on the show quite often. Introduction to Statistical Learning and The Elements of Statistical Learning. All right.
Nisha Iyer: Yeah.
Kirill Eremenko: Yup. Thank you very much for sharing those and also once again, thank you so much, Nisha, for coming on this show and sharing all your amazing insights today.
Nisha Iyer: Well, thank you so much for having me. I enjoyed talking with you.
Kirill Eremenko: Same here. All right. Take care.
Nisha Iyer: All right, thank you.
Kirill Eremenko: So there you have it. That is Nisha Iyer. Hope you enjoyed this podcast. My favorite part, my favorite takeaway from today was this whole idea of organizations more and more, with time, upskilling their workforce in order to be able to be data literate and to be able to work with data. It's something that's going to affect all of us and all of the people around us, everybody we know in a good way. An apt way that Nisha put it was like before, long, long time ago, nobody needed to know Microsoft Word. Then, with time, more and more people were required to know Microsoft Word. And today, you can't really get into almost any role or at least any corporate role without knowing Microsoft Word. It seems like a very essential skill for us now. So something similar is most likely going to happen with data science and data science is here to stay.
And you can get all the materials for this episode at www.SuperDataScience.com/167. There, you'll find the form that Nisha mentioned that if you fill out, you'll get their toolkit that they supply at the Data Society. You can also get all the show notes, the transcript for this episode and the links to the books that Nisha mentioned. And, of course, the links to Nisha's LinkedIn. Make sure to connect with Nisha and stay in touch. And on that note, we're going to wrap up for today. I look forward to seeing you back here next time. Until then, happy analyzing.