Welcome to episode #018 of the SDS Podcast. Here we go!
Today's guest is Analytics Industry Expert Jen Underwood
Listeners, what a treat I have for you today! We are joined by Analytics Industry Expert Jen Underwood as she looks back at the wealth of experience that two decades in data analytics brings.
Her long tenure gives her exceptionally strong foresight into what the future holds for the world of data analytics and she details the 6 megatrends she sees for the industry in 2017 and beyond.
Join us for a glimpse into the future in this exciting episode!
In this episode you will learn:
- Entrepreneurship (4:55)
- 6 Megatrends – Overview (10:35)
- Digital Business Transformation (18:49)
- Cloud Computing (22:17)
- Intelligent Analytics (31:27)
- Smart Machines (36:30)
- Internet of Things (45:57)
- Physical Plus Virtual World (49:30)
Items mentioned in this podcast:
- Data Preparation for Data Mining by Dorian Pyle
Kirill: This is episode number 18, with Analytics Industry Expert Jen Underwood.
(background music plays)
Welcome to the SuperDataScience podcast. My name is Kirill Eremenko, data science coach and lifestyle entrepreneur. And each week we bring you inspiring people 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.
(background music plays)
Happy new year everybody and welcome to 2017! Isn't this the most exciting year ever? We've got so much going on, the world is dynamically moving into the space of data, digital, data-driven technologies, and it's just so exciting. I can't wait to see what will happen this year in all those spaces.
And moreover, on that note, we've got a super special guest today. Jen Underwood is an industry expert in the space of analytics. She has over 20 years of experience, and what we talked about today is specifically that. We talked about the trends in technology and data for 2017. So Jen, as she says, constantly keeps a pulse on the analytics industry and is up to date with all of the technology trends of everything that's going on. And moreover, she's got a consulting business in the space, she does all sorts of different work around analytics and data science, she's also got a blog on data science. So this is a person that is definitely on the cutting edge of technology.
So in this podcast, you'll find out quite a lot. We talked about six megatrends that are currently happening in the space of technology, and at first you'll get an overview of what they are, and then we will dive deep into each one of them and I'm sure you will pick up a lot of new and valuable insights. Personally, I learned so much from this podcast. So I'm super excited and can't wait for you to listen to this session. Without further ado, I bring to you Jen Underwood, Analytics Industry Expert.
(background music plays)
Hello everybody, and welcome to the SuperDataScience podcast. Today with me on the show, I've got a very special guest, Jen Underwood, who is an Analytics Expert with years of experience. Hi Jen, how are you going today?
Jen: I am doing great.
Kirill: So tell us a little bit where you're calling from.
Jen: I am calling from hot Tampa, Florida, in the United States.
Kirill: Yeah, it's still very hot? We talked before about this.
Jen: It is! And it's supposedly winter, but I think it was -- and I don't know the Celsius conversion off the top of my head -- but 85 Fahrenheit.
Kirill: Wow, that's -- I don't know the conversion as well, but that sounds pretty hot. So yeah. It's funny, because I was in the US earlier this year, and even not remembering the conversion, you like switch over to Fahrenheit there, and then you're like, ok, that's hot, or that's a bit cold, but yeah, 85 sounds pretty hot.
Jen: It is hot, yeah.
Kirill: Alright, so it's great to have you on the show, and just for our listeners, could you tell us a little bit about what you do?
Jen: Oh, ok, great. So I've been in the industry a little over 20 years, it's going to be 21 years this coming May, when I graduated.
Kirill: Congratulations, that's a huge number.
Jen: You know, it is. And essentially, a lot of it's driven by passion. I knew I loved data, and it's taken me into different directions that I really enjoyed. So I started off implementing, meaning building, data solutions. Did that for about 15 years, and then I moved into more of designing solutions. So I designed solutions on the product team at Microsoft, but also helped them go to market from the technology, speaking technically to buyers. And now I'm what's considered an industry analyst, and I also do some market research, and I still have that creative spirit in me, so I'm doing some marketing as well. I always play with data, so the market research, as soon as I get the data and I'm crunching it, but essentially it's been a really different and unique path, but yeah, almost 21 years of this craziness! Yeah!
Kirill: That's crazy, yeah. And right now, your LinkedIn is very impressive. You've been in so many different companies, so many different roles, and all sides of analytics! It feels like you’ve covered everything. And right now you have your own company. Is that right?
Jen: Yes. And the beginning of my career wasn’t planned at all to move around much, but my husband was in the military and we were forced to move every 18 months. I felt like I was a permanent temporary employee. And at the time I remember thinking, “Oh, this is really scary and challenging and I’m always interviewing,” but then at the end of the day, it really allowed me to have many different experiences, and see all sorts of different ways that people structured their architectures or their reporting systems and play with different tools along the way. So from an ADD perspective – Attention Deficit Disorder, which I probably do have but I just joke about it – I like to learn so much that I think it was really a neat lifestyle.
And now I have my own company, I don’t think I will ever go back. I like to set my own direction and pick and choose per se the types of things that I want to work on, and I go after it. So instead of when you’re consulting and you get assigned a project or maybe you’re stuck in the same thing for a long, long time, now it’s like “Nope, I don’t want that situation. What is it that I really want to do?” and then I’m going to hunt down some group that does that and see if they’ll let me work on that.
Kirill: Yeah, gotcha. So your company is like a consulting firm for analytics. Is that correct?
Jen: Well, it’s interesting. It’s a mix. I would say it’s mostly market and product research at this time, where I’m helping vendors design products, understand the market. I do a lot of hands-on testing of these tools, which is really fun, and saying “I like these features. I don’t like these features. It’s missing these things that I would really need as somebody that loves data,” or I’d say “This is really not the flow of how I’d be looking at it. I really need to have the script at this point, or the output at this point, looking this particular way.” And that’s been really fun. I also help people write content and communicate effectively with other people that love data, because a lot of times when you look at some of the content that’s coming your way, it’s just fluff, or it’s meaningless and it’s just not effectively communicating with you. And I’ve really been enjoying that aspect as well. So it’s a mix of product and market research, and also a little bit of writing.
Kirill: Okay, yeah. And you have a personal blog as well, where I notice you have lots of very interesting posts on the subject.
Jen: That is my passion project. I don’t think people realize how much time that crazy blog takes. I would say easily each post is probably a full day. They usually start around 11:00 p.m. or midnight, and then 4:00 or 5:00 in the morning I’ll finish up. And I always think “Oh, it’s just going to take 2 hours,” and then it always takes about 6 or 7 hours to really create a good post. But for me, that’s my creative outlet, or that’s my creative spirit, that gets to shine. Because a lot of times in projects you may not have that ability to be creative or to get that side of the project. So it is my hobby.
Kirill: Yeah, I understand. It’s completely up to you, right, what you choose to write about, what you want to cover off in the next topic and so on?
Jen: Yes. And call things out in the industry. I think the biggest thing that I’m seeing right now is the media is a bit corrupt. I’m going to be open with saying that. They’re telling one side of the story on a lot of different things. The vendors certainly aren’t going to tell you everything that you need to know to evaluate solutions. So a lot of times, I’ll elevate topics that I think are really important for folks to realize. They may not realize it because they don’t have the technical background. Or just topics that are fascinating. A lot of times I’m predicting trends, so I see different trends in the industry. My most popular blog this year was “BI Wars”. It was a “Star Wars” themed—again, really creative, different --you’re probably not going to see that in any normal tech journal! But it was real stats, it was real information, and I’ll be darned if a month later, some of what I said came true, which has been really interesting. I didn’t think it would happen that fast. It’s also an outlet for me to call out vendors or to just keep the world honest with data.
Kirill: Yeah, that’s totally great, and I’ll definitely check out that “BI Wars”. That sounds like fun. And I’m glad you mentioned that you’re excited about predicting the trends, because that is exactly the reason why I reached out to you to invite you on this podcast. I would like to talk about the industry trends. Or actually, to be fair, I knew there was lots of stuff you could talk about, but during our conversation via e-mail, you mentioned that this is something that you’re very passionate about, and I’m super excited to talk about the trends, especially leading into 2017. This year is kicking off, and we want to know what’s happening, what’s going to be happening in the world of tech, and the world of data, in 2017, so I’d love to start that conversation. Could you maybe outline for us the main industry trends, and then we’ll dive deeper into each one of them?
Jen: Okay, so there’s a few that are what I would call “megatrends”. You’ll hear a lot about this. And the core theme around them is this concept of digital business transformation, and digital business transformation incorporates a few different types of movements that you’re seeing.
The cloud and the whole concept of cloud computing has really proved to be very compelling, and we are finally seeing the fears of trusting the cloud, moving to the cloud. The laws in different areas of the world are changing. The government, different types of agencies that historically prevented sharing of data and whatnot – all of these are contributing to a bit more cloud adoption. So we’re seeing that shift happen a little bit more quickly now. Amazon just had their big event and shared where those were. Gartner had an event this year and shared how far. And it was fascinating to me, in Gartner’s presentation, that 49%, almost half of CRMs were in the cloud. I went “Wow! That’s a lot.” So that’s one of the big pieces.
The other thing we’re seeing is intelligent analytics and that’s fantastic for this audience and myself included, and that’s taking things like the whole concepts of cognitive analytics, artificial intelligence, deep learning, predictive analytics, and prescriptive analytics that for years, I was hoping folks would get there. That’s finally starting to become key as data is seen as digital gold, and it’s the data and the information that you can extract out of data that’s really the game changer in this. Where everybody is pretty much equal in this cloud world, that’s really the differentiating—you know, analytics being at the core of that transformation. And when we do these, or when I give some of these sessions live, I show a video of the computer and it’s a robot that’s so humanistic that if she didn’t have this little patch showing it was a computer – her name is Sophia – it’s scary how realistic the machines are becoming. I’m not a sci-fi movie watcher, but honestly, I feel like we’re a few years away from the sci-fi world of “Is that a human or is that a computer?” Because on the phone, you can’t tell if it were Sophia answering the phone. And these chatbots that you’re working with, the Watson chatbots and whatnot, they’re so realistic and you wouldn’t know. So that’s really getting to be quite interesting.
We have what’s called the Internet of Things finally taking explosion. There has been a very interesting research around things will certainly be more than people in this world, and in fact they are already. So it’s these smart things and all the digital signals that they’re sending, and this whole Internet of Things becoming its own kind of entity and creature to navigate in what I would call—there’s this other piece of physical and virtual world, and that’s the area that I’m honestly most excited about, and it’s more just because I think—that thing was really called “third world”? What was that world, that there’s another like alternate world, or something? And folks would tool around and they would have other identities and whatnot. In this physical and virtual world, this is where you’re taking augmented reality and virtual reality, and you’re able to see and experience as if you’re right there.
And even with data I’m seeing it, being able to immerse yourself and dive into charts, look at them in all different directions and experience it. There’s a bit of simulator sickness that you experience when you start to experience data this way, or even some of the other types of virtual reality viewers. But it’s fascinating to see maybe different types of applications of these technologies from health care to helping people simulate real world situations to even just experiencing. I have a friend that’s disabled, and I showed her the Google Cardboard glasses with virtual reality and the Discovery Channel, and being able to experience going through a jungle that she never would have the opportunity to do otherwise. And jumping around and looking 360, it’s really cool. So those are what I would call the top trends. And we can certainly dig down deeper into lots of the other little ones that might be easier for folks to grasp.
Kirill: Thank you very much. That’s a great overview. Yeah, I’d definitely like to talk more about just that example of the virtual reality that you just mentioned. I totally agree on that. I myself just tried out the HTC Vive recently, and it feels so real. You’re climbing Mt. Everest, or you’re shooting a crossbow, or a bow and arrow. It’s insane how realistic it feels.
Jen: Yeah, it’s really cool. We just went—and I’ve had the Google one for a while, it’s very low cost. We bought the Oculus Rift, I believe is what it’s called. It was on sale, so we got that. It’s just a whole new way to experience watching and just interacting with entertainment that I’m certainly not used to because I’m not a gamer. Oh, that’s actually a good point. So gaming and this gaming industry is really influencing tech. And that is kind of fun and cool, but it’s also—you know, it’s just getting really interesting.
Kirill: Yeah, totally. But it doesn’t always have to be gaming, right? With those virtual reality tools, for instance, with HTC Vive, you can just switch on your desktop and actually create a virtual workspace and work on your actual projects but in a virtual workspace. How crazy is that?
Jen: Oh, that’s neat. Yeah. So one of the solutions that I was looking at earlier this year, I think it was called Datascape. They’re based out of London. And there are certainly other ones coming for the data analytics space, of being able to analyse your data in virtual reality. So, that’s been a really fascinating area that I’ve been looking at. But yeah, definitely not gaming. Where I’m seeing the most potential adoption, or we’re seeing it already, is early retail, new retail experiences being able to experience the product in 3D or imagine yourself in this cool car. Yeah, that industry is certainly jumping on it, but we’ve also seen—my sister is in healthcare, and she’s been doing what she calls sims, or simulators, for a long, long time. Certainly IBM Watson is kind of dominating that space right now. But all these different applications of training folks to getting a chance to see and experience, to entertain, and just looking at things from a whole new perspective of reinventing.
Kirill: Yeah, totally. There’s massive opportunities for application of these things. Okay, so just to sum up your trends, and then we’ll go into them, so our listeners can follow along, we’ve got digital business transformation, we’ve got cloud computing, intelligent analytics, smart machines, Internet of Things and physical plus virtual world. By the way, if anybody listening to this is not in their car, because a lot of our listeners listen to our podcast on their way to work, you can google and look for Jen’s presentations on SlideShare and maybe as we go through these, you’ll be able to get some additional information from the slides that you’ll find. All right, digital business transformation – could you tell us a bit more about that? What does that mean? What does that imply?
Jen: So the digital business transformation, and I have a session on this specific one – I have a blog jenunderwood.com and events – there’s a recording as well and there’s also a blog that details what digital business transformation means, because it seems really generic when you listen to it. But when I start to dig into it I think it will begin to make sense.
A great example, and a lot of times I like to use examples, is if you think about Uber in the concept of sharing resources that already exist, or some of these new business models that are very different than historical business models, but they’re relying on technology at the core and just changing the way that the customer experiences the solution or the provider. In this case it’s signing up to get a cab – in this case it’s an Uber. All sorts of different approaches to it. So it’s reinventing a business to a digital model and looking at what we would call “the customer journey”, and the customer journey being, you know, historically maybe you would go on a website and order a cab; or you might call a cab from a phone number, and then there’s a certain queue of cars that would be available, and they’d send it out, and tickets and orders. Well, Uber completely reinvented that and said, “Well, just through this little app on your phone, wherever you’re at, we can then send it to folks that will be nearby. All these cars are idle. You don’t have to then purchase.” I think in New York City I’d heard the last rate for being a cabbie was something like $500,000. That’s like buying a home. So it’s a mortgage to be able to be a taxi driver. Being an Uber driver, you just get to download the app—I think you apply, go through a background check, and you’re good to go. I mean, it’s really removing that barrier to entry, using technology to more effectively reach that audience and just altering the system. But the key is what we’re seeing happen in that space, is that whole concept of looking at the customer journey, what pieces can be digitized and automated or changed to be more flexible and leverage some of the assets that we already have? Those are some of the commonalities.
Kirill: Okay, gotcha. So take businesses or ways of doing things that existed previously and are not really aligned with where the world is going, and change them to get them up to speed. A great example with Uber. Same thing with Airbnb – don’t build hotels, use existing assets.
Jen: Yeah, exactly. Amazon is disrupting retail.
Kirill: Totally. We can keep going, right? Facebook doesn’t produce any content, but it’s the biggest content marketing platform in the world and so on. Alibaba also changes the way that you can start a business and take an idea from inception to actually bringing products to your customers very quickly. Yeah, that’s definitely something that we’ve been seeing. So you think this trend will just become even stronger in 2017?
Jen: I do.
Kirill: Okay, beautiful. All right, so we should look out for more companies like that. And if anybody has ideas for companies and how to change things, it’s definitely worth looking into.
Next one was cloud computing, and you mentioned an interesting statistic that 49% of CRMs are already in the cloud. That’s a huge number. That’s like half of customer relationship management systems. Is that across all of the ones that exist in the world, or is that across a sample of some certain industry that you’re looking at?
Jen: So, in I believe it was October or November—no, it was October – of this year, Gartner has an annual conference with the C-level executives in Orlando, Florida. And this year they talked about the stats on the Gartner audience and most of the Gartner surveys and audience and customers – those are large corporations. I would say you are not going to see a lot of the smaller ones. So when you talk about what’s the sample survey that they used, it’s their customers. And here are some stats from that, because I thought it was really interesting. The 2016 annual spend – and I do have a blog on this, by the way, it was October 19th, that’s when it was – 49% of the sales in CRM were already in the cloud. The next biggest one was marketing at 27%, some digital commerce, which makes sense. What isn’t quite in the cloud yet are things like manufacturing operations, so those are still computers probably in a warehouse somewhere. And what’s kind of in-between both of those at 20% is HR systems and procurement systems.
Kirill: Okay, a very interesting stat. So slowly we are seeing things moving to the cloud. And by “cloud,” we obviously mean just “off premises”. We mean data storage, like whether it’s on—for example, if somebody’s using Dropbox, that’s an example of your data being on the cloud. It’s not on your personal computer, you don’t carry it around with you all the time. It’s in the cloud and you can access it from other different locations. So just to kind of sum up for our listeners, what would you say are some of the advantages and disadvantages of the cloud? Like, why is it taking so long to take off?
Jen: This is really interesting, and I have some recent experiences with this. The benefits of cloud are the ability to rapidly deploy innovation—so even historically, when I was a product manager at Microsoft, in the past it would be every 5 years we might release an update, or then it became every 3 years. Then we were really excited and it would be one year. And the solution there was a cloud-based solution from Microsoft that I was on the product team for last year. We were releasing almost once a day.
Kirill: No way.
Jen: And Amazon said the same thing. They were having over a thousand releases in 350—how many days of the year are there?
Kirill: 365 days.
Jen: Yeah, exactly. So it’s more than one per day! And I thought, “This is just staggering!” It’s so compelling. And the other thing that was really interesting, and we’ve certainly seen it, and what cloud allows you to do is what Salesforce did to Siebel. So Siebel was on PRIM. They were the leaders in CRM. And this is for older folks. Older folks may get this but the younger ones may not. But basically Siebel was the CRM system and the one, the leader. Well, Salesforce came in the cloud, and they innovated very, very quickly and far surpassed them. So when you think about some of the ease – so some of the benefits as the speed of innovation, and I’ve personally experienced it – and now when people come to me and they have only on PRIM, I warn them that they can easily very, very quickly get surpassed from an innovative standpoint if they don’t have a cloud-based solution. I was anti-cloud maybe 2-3 years ago until I experienced it myself.
Kirill: Oh, really?
Jen: Yeah, and the reason why—there’s an important reason why, and Gartner confirmed this as well. It’s very hard to estimate the challenge of cost, you know, how much is this really going to cost me? And Gartner confirmed that really it’s very, very difficult, and there’s a lot of surprise bills that people are not prepared for. And when I think about, “Boy, it was just nice to buy this packaged software,” and the customers that I would work with, they would have a database maybe for 10-12 years where like, “Oh, gosh, you really need to upgrade every once in a while.” But hey, it works. We don’t need to fix it. So this model of constantly paying—you’re paying a lot more for cloud, you really are. So the downfall is certainly that.
And right now what I’m seeing, at least in some of the early adopters of cloud, is these shocking bills. There’s a customer in Tampa, I can’t say who they are. They just laid off 40 people because they had a $2 million over [indecipherable 27:16] bill that they weren’t expecting. And I’m like “Oh, my gosh!” You really need to be prepared and proactively very, very closely monitor your cloud costs because it will alter your business. And that’s what we’re seeing is folks having to adjust their whole business model. We’re seeing companies like ClickTech saying “We need to reinvent ourselves for cloud now.” So they went from being a public company to a private company to give them the chance to restructure and reorganize their business. So there’s a couple of downfalls you need to be ready for. The good part is it’s easy and you get a lot of new features very quickly.
Kirill: Yeah, and that’s some very interesting insights to the company going from public to private just to get ready for the cloud. And with your comment that it costs much more to be on the cloud I will agree and disagree. On one hand, it probably does when you go to something like Amazon, like when they have this EC type of billing system, where they’re just billing you per hour or per minute and so on. It can really accumulate very quickly. And it is an expensive thing, because you’re paying for a premium, for the service, the security, the updates, that agility, the speed. It’s something that they understand, these providers, that they can charge for that premium.
But at the same time, imagine a huge company with 6,000 employees that has servers, and then all of a sudden they need to upgrade. They have to spend like $20 million just to upgrade the servers. It’s a one-off huge cost. Even though cloud might cost a lot over a long term, so every month you’re paying huge bills, if you manage it properly and it’s predictable, it’s manageable. But when you don’t have any costs for 5 years and then you have to upgrade for $20 million, that’s a huge cost in one year, and that can really sink companies sometimes.
Jen: That’s interesting you say that, because I remember when one of the cloud data warehouses came out. My peers in the network that build a lot of data warehouses, they were ecstatic, and the reason why is—and you see a lot of excitement around these cloud data warehouses—well, one of the reasons was because then they can take all the money that you would spend in the hardware, and now put that toward services. And that’s how my friends make a living, is by selling their data warehouse building services, and they’re like “Oh, we have more money to spend! So this is great!” Yes, you absolutely have a point there, and it was confirmed even by the reactions of my peers.
Kirill: Yeah, all right. Well, thank you very much. That was a great dive into cloud computing. Oh, I just wanted to mention this very funny thing I learned recently about cloud. We constantly think about cloud as this celestial thing, like “up in the cloud”. But actually when you think about it, the cloud – what is that – that’s just somewhere in a server, somewhere far away, but how does information travel across continents? It travels by cables—
Jen: In the ocean.
Kirill: In the ocean. So the cloud is under water, if you think about it.
Jen: Yes, it’s absolutely true. And some of the servers are in the ocean to be cooled by the ocean, which is—I have some mixed feelings on that because I’m an environment lover. Yeah, I’m not sure if I love that or not, but yeah, there’s even servers in the ocean too.
Kirill: That’s great. Do you have a blog post about that?
Jen: I don’t.
Kirill: You should write a blog post about that! That is so fun.
Jen: Yeah, that’s a fun concept to talk about, the cloud being under the ocean. Because the ocean is very mysterious to – I mean, even really, to the scientific community. They’re constantly finding new creatures and new things in the ocean.
Kirill: One day they’ll find some servers there.
Jen: It could be a fun theme. If I do, I’m going to definitely send you that link.
Kirill: Okay, sounds good. All right, moving on to the next one: Intelligent Analytics. And here you mentioned that finally – you’ve been hoping that the world moves to this for quite some time now, and finally we’re moving to machine learning, not necessarily artificial intelligence, that’s the next one. So machine learning, different modelling techniques, and really upskilling the way we use data, because data is the digital gold of this era. What else can you add to that, and where do you think this trend is going in the next year? Or in 2017?
Jen: So, the Intelligent Analytics is a fun area to see evolve. Essentially we’re moving from the maturity level of analytics, of historical reporting, and now we’re beginning to see more forecasting and predicting, but the technology again is moving so fast that various vendors in the space are automating the predictions, and going through all the combinations of variables that historically someone might do by hand, and look for the correlations, or they’d look for the variables for data science, the most information gain, if that means things to folks on the line. And looking for those interesting patterns within the data, and transforming variables to find what I would call the art and science of data prep.
Well you have vendors right now that are automating those processes and automatically creating predictive models and showing you the results, and not just predictive but also prescriptive, making recommendations of what to do based on the scenario that you have. Now, these are things historically—you know, you only had a few people maybe in an organization, a large organization, that could do it, we’re really talking about 1-2% of the population that’s able to do prescriptive analytics, that even understands what optimization and simulation even means. That’s a very rare skillset to find. But we’re seeing some of that being automated.
There’s a lot of scepticism that I have around that, and so do some other folks. There is some good and some bad. The good is that some of the dirty work is being automated, and you’re getting some instant things. When you run some of these automated insights, you’ll within a minute or two see some of the things that may have taken you much longer to find, if you were to find it. It’s unbiased, it’s going to run across everything, so you may not realize that you have a data science bias when you’re applying it, but subconsciously you might. So you’re removing the bias in the analytics.
One of the fears that I have with this is it’s garbage in/garbage out. A lot of times the predictive model needs to really reflect the business process you’re trying to predict, and you may not know all the variables in that process where it’s very iterative in nature. And one of the examples I had was for an insurance company. We’re trying to predict churn of their reps that come in to sell because it’s very expensive for this company to train them and then find out they weren’t a good fit from the get-go. So what we were doing was creating a model. The first iteration of the model says “Don’t hire anyone from New York City.” That was the output! So we take it as the techies to the business, and the subject matter expert says, “Well, that makes sense. There was a law that changed.” We went “Oh, we never thought to include the regulatory laws for that in the model.” So it’s things like that where I say, “You know what? I just don’t think every single thing can get automated, and we do need to be aware of the limitations of these tools.” But to a naïve or an uneducated buyer, “Oh, I could see somebody that really does data science. What’s taking you so long? What’s the problem? I just saw them pick the easy button. You should have an answer to me by now.” So I’m hoping that we don’t have a bunch of bad decisions, and we don’t have missed expectations or unrealistic expectations for folks that really do the true art and science of data science.
Kirill: Yeah, gotcha. And I totally agree. Even though machine learning and all these new modelling techniques are picking up, there is still a gap that only humans for now can fill in, where you’ve got to apply that domain knowledge, you’ve got to apply that creativity, you’ve got to do that. There’s some research that you cannot get from the data. You go and sit down with people working in the business and you find out insights like this, you know, that there’s been a legislation change and that’s something that’s not part of the input for the machine learning algorithm, and therefore it’s only a human that could have adjusted that. But with time, probably 10 years down the track, machines will be doing everything, I think, but for now—
Jen: It is scary, yeah.
Kirill: Yeah. For now, I totally agree that people should have realistic expectations. That if they want a great model, something that’s really working, it does take some time and some creativity to come up with something like that. All right, so that’s Intelligent Analytics.
And next one was Smart Machines. And here you talked about robots and machines like, for example, IBM Watson and so on, that sometimes you can’t even distinguish whether it’s human or it’s a machine. So what is your prediction for the space of robotics and robots and them integrating with our roles? Do you think it’s going to start in 2017?
Jen: Well, I think it’s already happened, and it’s happened in different parts of the world faster than at least it has in the United States. If I look at Asia and countries like Japan and China, they’ve had robots already. We do see them in the United States as well, in things like warehouses for picking and bagging and tagging inventory type systems. But what we’re seeing happen in Asia right now are robots that can bring you food service when you order food service from a hotel. When you come into a store, you can interact with them and ask them what is the type of—I think it was Nescafe that was using this – again, retail getting creative with experimenting early with technology.
And experimenting, by the way, is a theme that I’m seeing in a lot of digital transformation, and a lot of playing with some of these technologies to figure out what works and what doesn’t, and the whole concept of “let’s experiment and fail fast” and just try new and different things. It’s actually being rewarded more so now where it used to—when I began my career, you didn’t really want to take risks too often. A lot of people were afraid to do that. And now risk-taking is almost encouraged because we’re trying new things on the bleeding edge. But going back to Nescafe, they experimented with I think it was Pepper the Robot in different stores to help consumers direct them to the coffee they might like. So certainly we’re seeing a lot more of that. It’s the cost of these, making these computers and these robots. In Silicon Valley I’ve certainly seen a few running around in parking lots, to try and share with people how far it’s come. But the cost is still too prohibitive for it to get adopted widespread right now. But I do see that we’ll see more, and probably more from larger organizations.
Kirill: All right. Yeah, I totally agree with that. Thank you so much for the overview. And my natural question here is something that I’ve also been pondering myself, is what will happen in the world when these robots start replacing a lot of the manual labour that we have. So when people start losing jobs to robots – like even a coffeehouse is starting to use a robot, or a hotel is starting to use a robot to deliver food and room service, what will happen when mass population starts losing jobs to robots? What is your prediction for that?
Jen: Well, it’s certainly scary for the next generation. I have to say I love to be positive about certain things, and in this case I can say I’m uncomfortable with the answers in some of the things that I see propaganda-wise in the media. So I’m hearing a lot about universal base income that I never heard about even 3 or 4 months ago really being pushed in the media today as jobs are being offshored, or there’s different types of trends with this automation. I think there will be a lot less jobs. The estimates are there will be 60% unemployment. And it’s being sold to the population right now in the media as “Well, this is an opportunity to live more and to enjoy yourself.” And it might be. You know, I’m a workaholic and I don’t know what to do sometimes on my time off but yes, I would love to do more with animals and nature. You only live once, and certainly looking at a computer screen every single day of your life is certainly not healthy. So I think there is going to be a balance that yes, we’re going to enjoy our lives more, so on the good part.
The bad part is our systems for mortgages and for loans and for different types of things need to change with this and need to adjust with these changes. The other thing then, I think, is looking at different types of employment, so I think what we’ll see is a renaissance of trades where it was the Oxford Martin School – I believe they might be in England – did a study on the probability of computerization and what jobs would and would not be around. And the types of jobs that would be around – they have an online calculator, it’s quite fun – essentially, the types of jobs if you have to crawl in small spaces, or interact and negotiate with people, if you have to care for people, those types of jobs will be around. The types of things that won’t be around are services jobs that we’ve really—you know, we went from being manufacturing to service, so this is the fourth revolution, the fourth industrial revolution, this whole digital transformation.
The World Economic Forum is another really good site to track what are the trends and what are policy makers around the world doing about this problem to help with this employment. Because now what we’re seeing is the services jobs going away. What jobs are left? So health care, taking care of people, negotiating… I think there’s going to be a whole different type of job. And when I say a “renaissance of trades”, you know, being able to build bridges might be really cool again, or doing things that require that person to be onsite and to deal with location issues because it’s hard to get a computer in certain types of scenarios to crawl under a house or to go into certain locations. Those types of things may be new. So we’re really heading into a bit of an unknown altogether really, all of us are.
Kirill: Yeah, totally. Like you say, very scary and exciting at the same time, so hopefully our policy makers and governments can keep up with the technology and make sure that it’s for the best of everybody at the end of the day.
Jen: Well we are seeing already in the artificial intelligence space the largest companies getting together and saying “We need to have some ethics in data science.” And I believe that just happened maybe a month or two ago. So we are seeing folks certainly talking about these issues, which is good. At least they’re not hiding from them. They’re addressing them, they’re bringing it out into the open to have conversations. It’s a matter of being aware of what’s happening. So you can only plan so far in advance, but it is certainly something to be aware of. So if you’re going for a law degree or something, the lawyer jobs may not be around so just be aware of that.
Kirill: Okay. And can you elaborate a little bit more with this ethics in data science, what did they talk about in that conference?
Jen: So, essentially what’s happening right now is the concept of — and you’re going to hear a lot about this as we talk about trends and what you’re going to hear about in the next few years. This has really started to surface with Salesforce, Marc Benioff, being upset about the LinkedIn purchase by Microsoft, and the control of the data because the data is key. And that’s why you see all these apps for free. It’s the data. So we’re going to see a lot more about “When you give your data away, did you agree?” So, I certainly did not agree to give my career information to whoever bought – it was Microsoft in this case – I didn’t agree to do that, maybe I’m uncomfortable with that. It’s, “What rights do you have with the data itself?” Facebook has lots of controversial uses of the data, as probably does Twitter and some of these other accounts as well. So things on how are we using the data, what rights are being given away, do you understand the privacy laws? Is it being, I think, on the policy level — and I felt this was really interesting recently: the White House in the United States has a paper on bias in data science and ethics in data science, and in their case they’re saying when you look at setting credit scores or making loan approvals and these types of things, making sure the data is not biased towards one specific demographic or another one, and making sure it’s really balanced and not skewed essentially. So there’s a lot of different angles of it, but it’s finally being discussed.
Kirill: Yeah, very interesting questions are being raised. And it’s totally a new space that’s emerging as a consequence of how quickly and in which direction the world has been developing in terms of data science, so it would be interesting to hear more and see what comes out of those conversations. All right, now we’ve got two more trends left, so Internet of Things is our fifth trend. What is the Internet of Things and what are your predictions for 2017?
Jen: So when I think about the Internet of Things, I think about all these little sensors that we have sending data, and it can be something as simple as it would be a sensor on a package that might be being shipped and you can tell immediately where it is and where it’s at in the shipping process. It might be a sensor on a vehicle saying that you’re speeding being sent to your insurance company. It might be a machine somewhere. I think one of the most fun gadgets is one of the Amazon buttons that you can program yourself and just click the button or you’re even [indecipherable 46:43] up with voice, being able to talk and order things.
So you have things sending information for you, or finding information for you. And the amount of data that things generate is beyond exponential. It’s just ridiculously huge. So there’s new architectures being designed that can just ingest and handle, and a lot of times it’s very periodic heavy loads. So think about maybe when you have a massive sale on—in the United States, at least, it’s the day after Thanksgiving we call it “Black Friday”. And then there’s “Cyber Monday”. And there’s massive loads of folks grabbing those great deal while they last. So your architecture needs to be very flexible and just bring in data in its raw format and just be able to ingest it. It’s called “ingestion.”
And it’s a very different design pattern from when I started in the industry and we would literally ask for requirements and we would design how the data would come in. So we might bring it in and store it somewhere, but then we would be very picky about what we might take and pick out and put into a data warehouse format to use. It’s very different now, it’s just “get it all in there”. And then we have really cool technologies. Amazon just announced Athena, that you can just query the raw data in S3 as is, and schema-on-read – just read it and query it and interact with it. No modelling, no database. I’m like, “Holy smokes!” No ETL at all, no data prep. I’m like, “Wow! This is crazy.” So it’s really getting interesting, the design patterns. The design pattern changes, just the way data moves changes and this concept of things and being monitored everywhere is also quite interesting.
Kirill: Yeah, totally. Okay, and you think that this trend will only strengthen in the coming year? You said “beyond exponential”, so we should expect even more data, even more—what are the three V’s of big data? Velocity, variety and volume, right? So you should expect that to just increase in 2017.
Jen: I do. And we’re seeing that, even with the consumer. I literally have like a Raspberry Pi gadget in here, and you’ve got people that are monitoring their garden. When they travel—they’re travelling consultants that monitor their garden with a Raspberry Pi. I always wanted one of those easy buttons so I thought “Well, that’s neat. I just want to play with it. I just want to be able to have my own button where I push a button and all these cool things happen with it.” So the technology is also being distributed to consumers quite easily so it’s accessible.
Kirill: All right, thank you for that, and we’ve got one more to go: the physical plus virtual world and all this VR augmented reality. Very exciting things! Where do you think that’s going in 2017?
Jen: So I see that first and foremost coming to a lot of folks when Santa comes this year. I’m serious. So what I see is a lot more folks playing, they get their viewers—the viewers are less expensive. They used to be $600. We got ours this year for $70. So it’s really very realistic with pricing now. The Google Cardboard was $20. What I think is now we have enough content in the library where you’ll see the consumers this year enjoying it, playing it, looking for it – we’ll probably see a lot more marketing at the retail, the National Retail Federation. I would expect to see more displays and more ways to attract from a marketing perspective and awareness perspective these campaigns with VR to experience their products in new and different ways.
I also expect health care analytics. And I was surprised – I did some market research in this area recently – just how pervasive it was in my particular network of analytics professionals. And what we’ll see there is the ability to use virtual technology in different industries like health care, like working with neural electrical equipment or different types of very difficult situations that you want to be able to do. It’s called simulations today – to simulate that environment. So I would expect to see much more there.
I’m not sure how much we’ll see in data. I certainly am very optimistic that some of the solutions will come to play. What I can see so far in the data visualization space is I’ve seen some fun demos of experiencing the stock markets, and it felt like you were on a rollercoaster literally. It was really neat and it’s fun, but it’s hard to synthesize because the general user, or the general audience, is not used to experiencing data by riding on a rollercoaster. And I looked at another one. It was the Salesforce example with Oculus Rift and I said “I don’t know what on Earth I’m looking at, what’s good and what’s bad?” So there’s a lot of learning. What I think we’ll see more of is, if anything, experimentation, also folks learning how to design user experiences so you don’t get sick, because we’re seeing that happen already in the mainstream space, and we’ll just see a lot of fun campaigns really this year.
Kirill: Yeah, sounds exciting. And for those out there listening, definitely pick up one of these devices, especially if they’re now cost effective, you know, $20, $50, just like I do it sometimes. Even though I don’t think I’m interested in something, some piece of technology, if I can afford it, I will still purchase it just because I want to experiment with it and just to be on that technological trend, to know where the world is going, because that might give me some ideas. I might come up with a business idea or I might come up with an application idea out of it. Like Jen said, experimentation is an important part of innovation. And especially around this festive season, why not pick up one of those devices and see what ideas you can come up with.
All right, wonderful. Thank you very much for going through all those six trends. I’m sure there was a lot of value. I myself learned a lot from all of these examples and from all your valuable experience and I’m sure our listeners will too. Coming to the end of this podcast, I’d like to ask you to share some ways our listeners can contact you and get in touch with you and maybe follow you or maybe even connect with you over the different social media and different resources that you have. What would you say are some of the best ways to get in touch with you?
Jen: Well, thanks for asking and thanks for having me on the show today. The best way to contact me is I have a website, jenunderwood.com. It’s my blog. There’s contact links on that blog so that you can fill out a form and just reach out to me. And I do get those and I usually answer fairly quickly most of the inquiries that do come in. If you want to follow me, I have a profile on LinkedIn. Google+ —I have a profile there as well. And on Twitter. I’m very active on Twitter. My – I guess you call it “tag name” or something – is @idigdata. So it’s an easy one.
Kirill: Gotcha. Yeah, I can attest to that. Jen is very active on Twitter. Some very interesting insights being shared there, so definitely, if anything, definitely follow Jen on Twitter, @idigdata, and get the latest and greatest of the technology updates. One more question I have for you today, Jen, is what would you say is your one favourite book that you can recommend to our listeners that will help them become better data scientists?
Jen: It is my all-time favourite. It’s an oldie but a goodie, is the Dorian Pyle “Data Preparation for Data Mining” book.
Kirill: And what do you like about the book?
Jen: It really goes into what I would call the art and science of preparing data, and that’s the magic sauce in a model, is how you would transform the variables. You know, sometimes even finding a pattern that wouldn’t even necessarily—maybe it’s the null values or it’s the records that weren’t filled out. That’s a variable. It helps you think about how to look at data a little bit differently for predictive modelling in general. So it teaches you how to think and approach data prep. Things like the skew and the bias I talked about that you won’t realise maybe that you have in your sampling techniques. It’s really a classic that’s taught in most – or at least used to be taught – in the universities, I don’t know if it still is, but it’s probably one of the best ones that taught me just the magic and how important that piece of the process is.
Kirill: Fantastic! That’s some great advice, and I totally agree. Data preparation and the way you think about your data, about your data problem and about how you’re going to tackle it are very important. So go ahead and check out that book: “Data Preparation for Data Mining” by Dorian Pyle. Thank you very much, Jen. It was a pleasure having you on the podcast and I’m sure so many people are going to pick up so many valuable insights from what you have shared today.
Jen: Thank you for having me.
(background music plays)
Kirill: So there you have it. How exciting was that! So many different technological trends and each one of them has so much substance to it. It was so interesting talking about this with Jen. I definitely picked up so much from this podcast myself. Probably the biggest insight for me was the thing that we talked about in digital transformation when Jen mentioned that it’s about reinventing businesses to a digital model and she gave an example of Uber, we talked a bit about Airbnb and all these businesses that are doing same things that used to be done before, but they’re doing it differently. They’re doing them digitally, they’re disrupting their spaces. You know, I’ve always thought about these things and I’ve known—everybody knows about Uber and Airbnb and all those companies and the way they operate, but I’ve never thought of it that way, that they’re actually doing something that used to be done a lot of the time before. It’s something that’s been around for a very, very long time, they just changed the way it’s done. They didn’t invent something new to be done. You know, hotels have been around for ages, and the taxi industry has been around for ages as well. They’ve just invented a new approach to the old problem. So it’s just a different way of thinking that I think I’m going to be a bit more open to now, or I’m actually going to encourage myself to think about things I see around myself in that way. But at the same time, everything on this podcast was super valuable, and I’m sure you’ve had lots of different takeaways and I could probably make a huge list of the things that I’m going to take away from this show.
I hope you enjoyed it, and if you did then you can find the show notes at www.superdatascience.com/18 because this was episode number 18, and there you’ll find the transcript for the episode, any resources that we mentioned, including a link to the book that Jen mentioned. Also you’ll find the links to where you can contact Jen such as LinkedIn, Twitter, her personal blog, her company website – there’s a lot of links to remember here, so definitely check out that page. And I highly recommend that you do follow Jen on Twitter. Her handle is @idigdata. The insights that she’s sharing are invaluable and they’re sometimes funny and they’re sometimes exciting and they’re always about technology, they always will help you keep up-to-date. And on that note, thank you so much for joining us today. I look forward to seeing you next time. Until then, happy analyzing.