New DataFramed Episodes

Hi, I'm Maggie, VP of Curriculum at datacamp, and I know firsthand how overwhelming it can be to keep up with the rapid changes in data and AI. But staying ahead doesn't have to be hard. That's why our curriculum team has curated the best learning content to help you and your team stay ahead, covering everything from data and AI literacy for busy executives to cutting edge generative AI for developers, and everything in between. Whether you're upskilling yourself or your team, we make learning practical, engaging and flexible. It just works. Power your data and AI transformation journey on datacamp.com.

Hi everyone, this is Adele, a data science educator and evangelist here at datacamp. I'm really pleased to announce that the Data Frame Podcast is back. Over the past few years we've seen the maturation of data science and machine learning and have seen many of their promises come to fruition. We've seen awe inspiring results from language models like GPT3, progress on self driving cars, and huge gains in computational biology with use cases like drug discovery and protein folding. However, most importantly, we've seen a much bigger drive to democratize machine learning and data science with the emergence of better, more inclusive tooling, digital and data transformation mandates and upskilling mandates across a range of organizations and industries. However, there is still more to be done and a lot more to uncover in this iteration of Data Framed. We are really excited to showcase the latest thinking in all things data and how it's impacting organizations across the world. From best practices leading data teams, to the challenges of aligning machine learning to organizational value, to the importance of raising data literacy and developing data culture, we'll uncover and demystify insights on where the field is headed and how to scale its impact responsibly across the organization. It goes without saying as well that none of this would be possible without the support of Data Framed listeners and Hugo Baun Anderson's role in humanizing data science with the previous iteration of the Data Framed Podcast moving forward, Data Framed episodes will be released on a bi weekly basis, that is every two weeks, and we're excited to share the insights of data experts from a variety of backgrounds and industries. However, before letting you go, I'm really pleased to be joined by datacamp's CEO Jonathan Cornelison for for a quick chat on where he thinks data science is headed and his vision for the future. Hey Jonathan, it's great to have you on to discuss all the exciting developments in data science and education.

Hey Adele, it's great to be on. Thanks for inviting me you're someone who's.

Been operating in this space for a long time. And I think both of us can agree that data science is becoming widely accepted as something not only experts need to know, but as a methodology and mindset for solving problems. This means both data teams have become much more strategic for organizations and data science has become more democratized. However, we're definitely still well away from full democratization. I'd love your insights on what you think are the major challenges we're facing to reach full democratization of data science.

Yeah, sure. So at a high level, I think there's two main challenges. The first challenge is something that's very personal to me. It's the lack of data skills. And there there's still a supply demand imbalance between the number of people that have data skills and the number of people that should have data skills. And you know, this shortage really exists across a spectrum. There's a shortage of data scientists, there's a shortage of data engineers, marketing analysts, financial analysts and so on. And for most organizations, the lack of skilled talent is really one of their biggest bottlenecks. And to resolve that, you obviously need scalable and affordable education and training. Now that's the first challenge. The second challenge is that in my opinion, there's a lot of room for improvement in terms of the tools that data professionals use. And we've seen in the past decades we've seen a tremendous shift from legacy software such as SAS SPSS to Python R and other open source tools. And the adoption of open source tooling has really pushed up space forwards. But I would argue that regardless, the tooling for data professionals is easily lagging that of software engineers by five to 10 years. What I mean by that is that we're still missing cloud based, collaboration focused and enterprise ready tools that data teams use. That's not how most data scientists work. They still work very much on their own machine, typically with open source software. And you're starting to see shifts and a change there. But I would argue we're in the early phases of that.

So let's unpack these challenges. Can you dive deeper into your thinking around the lack of data skills?

Yeah, sure. So maybe start with the good news here. I think in the past years, hundreds of thousands of people acquired data skills and that's super exciting. And as I'm sure you'll agree, education and training is the solution to the lack of data skills. But I think the core problem is that a lot of the options people have think about traditional degrees, boot camps, they're still very expensive. It's not uncommon, especially in the US that you pay 50,000, $100,000 for a data science master, or even if you do a boot camp, that will easily be 5k, 10k, 20k, even 30k. And that's a lot of money for a lot of people. And it basically means that education is not accessible to a lot of people. And I think this is especially relevant given the impact the pandemic has had where a lot of people are kind of struggling financially, especially the people that were already struggling. And so in that context, I think it's super important that we continue to invest in really democratizing the way people can gain these skills. Now, if you zoom in on organizations and unpack the challenge there, I think it's a little bit more nuanced. There's different types of problems organizations face on this front. The first problem is that they're really struggling just to attract data talent because there's not enough people out there. And then second of all, they're often struggling to upskill existing talents as well. And it doesn't help that historically, I would argue, a lot of organizations have underinvested in kind of data skills, and now they're facing additional challenges. For example, how do you certify the kind of data science skills people have across the organization? Because right now, compared to the cloud space, for example, where there are very widely accepted certificates, that doesn't exist yet in the data space. Now, all of that being said, I am optimistic because what you're starting to see in the data space is similar to what happened to software engineering services space. Right now, I think most employers that hire software engineers really only care about the skills people have and maybe kind of the cloud certifications they have. And it's much less about, do you have that fancy degree from a certain type of school or a certain type of bootcamp? And you're definitely starting to see that shift that happened in the last, say, 10, 15 years in the software engineering space. You're starting to see that shift in the data space, which in my mind means that we're on the right track.

I couldn't agree more. As a data science educator, one of the reasons I joined EdTech is witnessing firsthand the power of democratized education. This is because through MOOCs, I was able to get acquainted with data science, and it provided me with the ability to get a graduate degree and get certified. Granted, I was privileged enough to pursue a graduate degree, and not a lot of folks can do that. And this is why democratized certification will be so important. This is especially true with the effects of COVID How do you see the data education space and the democratization of data skills further evolving to address these needs?

Yeah, I think the online education space now is already at the point where it's becoming a real alternative to traditional education as well as a complement to it, because a lot of online education is used in traditional educational contexts, which is great. What I'm personally also really proud of is if you look back to the last year, a lot of players in the online education space really stepped up during the pandemic. So think about Coursera, pluralsight, Addicts, codecademy, datacam. We all gave free access and scholarships to people in need. And so to zoom in specifically on what Datacam did, we launched our Datacam Donates initiative, specifically focused on empowering disenfranchised and disadvantaged communities that were affected to Covid. And we essentially gave free subscriptions to those groups. We still have that initiative, and we're actually scaling it to hundreds of nonprofits, and we're giving away hundreds of thousands of scholarships. And in my mind, this really shows the beauty and the power of online education because the marginal cost is very low. That enables you to really level the playing field and democratize access to data skills.

Now, one thing you mentioned is the upskilling imperative for organizations today. Can you walk us through your thoughts on why upskilling for data science and data literacy is important for organizations?

Yeah, I think it's important for several reasons. And I think the most important thing is that as an organization, you want your subject matter experts in every domain really to be data literates, because, number one, it allows them to ask the right questions. And number two, they can start answering their own questions. Number three, you want subject matter experts as well as the data team to speak the same language, because if they don't speak the same language, you're never going to really create any exciting data products or have a functioning AI strategy. And maybe the last thing I'll mention is that in the foreseeable future, data teams are going to be understaffed. And so any strategy that kind of helps alleviate a burden on data teams and especially centralized data teams, I think any strategy that does that will be a winning strategy. And kind of improving data literacy across the organization definitely does that.

Now let's unpack the second challenge, which is the state of data tooling today. Can you walk us through this challenge and how you think it's affecting data teams?

Sure. And I'll start by taking a Step back here. So let's think about how does data flow through a modern data driven organization? It starts with data collection. So data is collected from raw sources, website traffic clicks, billing data sensors and so on. And then ultimately it gets centralized in a data warehouse or a data lake, it gets cleaned by modern ETL tools and, and the data engineers work their magic and they make sure everything is in a clean, organized, digestible format. So that's all great. And then the final step is all of that gets turned into insights or actions by data analysts, data scientists and subject matter experts. And so if you think through that flow, where we've seen a lot of innovation in the past 10 years is in the bottom of the stack that's gotten pretty mature. So if you think about Amazon, Redshift, Snowflake, dbt, what that infrastructure layer made possible is that you can now have massive amounts of data in the cloud. It's fairly secure, it's easy to query. But where I think we'll see the next layer of innovation is kind of in the top of that stack because data teams still don't have the same level of mature tooling that software engineers have. For example, as I mentioned, and I want to be clear that there has been a great amount of open source innovation, but what I'm specifically talking about is kind of that translating into mature cloud tools focused on collaboration for data teams. And I think that area has been underinvested, which was natural because you first need that kind of bottom of the stack to mature and only then is there an opportunity to kind of build built top of the stack. And so I think we're going to see a ton of innovation in kind of tooling for data teams in the next five to 10 years. Anything from simple things like version control, quality control, real time collaboration, easy data access for team members who use R and Python. I think there's a lot of room for improvement in all of these areas.

So how do you think the immaturity of data tooling hurts data teams and how do you view the importance of tooling innovation, especially at the insights level, to further drive data democratization?

Yeah, in my mind it's really essential. A comparison maybe is think about the impact GitHub has had on innovation in the software engineering space. So I think we're about to see very similar dynamics in data space. And Obviously I know GitHub is being used by data scientists at the moment, but was it really designed for them? Was it really designed to share reports, dashboards, insight? Obviously no. If you think about the outputs of a data professional, that's not just code. It's always a combination of code, data, a presentation layer, and I think there's tools specifically designed for data scientists and data teams that we'll see in the next few years. We're seeing some innovation in kind of the notebook startups. That's, in my mind, the first iteration of that, and I think we'll see future iterations in the next few years as well, which is super exciting.

When looking at these challenges and if you want to work back from the solution for these problems to be solved, what do you think is the North Star for the data education space and the data tooling space and how do you see them playing out?

Sure. So I think the way I think about this is as an individual, I want it to be very easy, fun, affordable to gain data skills. I want it to be super easy to signal that I have those skills to an employer through certification. Then I want it to be super easy to get a job and to collaborate with team members. If you look at the dream end state from an organization point of view is that for them it becomes way easier to attract and upskill data talents to ensure that they're certified. And this is kind of a standard within the organization around how skills are measured. And then finally, you want to make sure that your data professionals and data teams have kind of the collaboration tools and the productivity tools that allow them to be productive. And I want to be honest here as well, implicitly. This is a shameless plug for what, what datacamp is working on right now, because we are rolling out a certification product and career services to individual learners. And we're also launching a collaboration product for data professionals.

That's very exciting, Joe. When can learners start getting certified and how can they get started with datacam's new Workspace product?

So right now, both the certification as well as the Workspace product are in beta. We're rolling it out gradually to the entire user base. And if, if anyone listening to this is interested, Definitely go to datacamp.com and check it out.

All right, that's awesome. Thanks for the insight, Joe. I hope you enjoyed this brief discussion with Joe and I'm excited to see where the certification and tooling space are headed for individuals, organizations and data teams. If you want to dive deeper into what datacamp is up to, make sure to check out the show notes. Also, if you're listening to this before April 30th, make sure to take advantage of free week, where everyone has unlimited access to all of datacamp's content library. Check out the show notes for more details as well. In the next episode of Data Framed, we'll be talking to Amin Rashauriki, principal Scientist at Nvidia and the former Chief Analytics Officer of the City of New York, on the state of data science and government agencies and the unique opportunities and challenges data teams face in this space. I hope it will be useful for you, and we hope to catch you next time on Data Framed.