#201 Graph Databases, Deep Analytics, And Change Management: The New Data Frontiers With Peter Kokinakos, COO of MIP

Graph databases are powerful tools in analytics, but they are an often-misunderstood innovation. As they hold the relationships between data as a priority, they are an invaluable tool for modern, heavily inter-connected datasets.

In this episode of Data Futurology, we explore graph databases with Peter Kokinakos (pk), the COO of MIP. They have been conceptualised for around 18 years, but it is only now that the computing power has started to catch up to allow graph database projects to come to fruition.

MIP is right at the front of delivering these capabilities to their customers. “It’s becoming a real product,” Kokinakos says in the podcast. “All of a sudden we’ve got the capability of delivering these really intricate kinds of analytics for complex relationships.”

Kokinakos, who will be speaking at the Advancing AI Sydney summit in August, further outlines the additional value that data scientists can get out of data relationship value in comparison to the data value. Delivering this value requires some change management to take advantage of because, as he says, “instead of just double clicking on something and drilling down the level, you can now actually drill down by the relationship.” However, once that change management process has been completed, the ability to be able to interact with customers on the basis of interconnected relationships rather than single data points is compelling.

Change management is a challenge for many organisations and data scientists – anything new is always going to have some resistance. This is why MIPS runs The Data School, and Kokinakos explains in detail the value that adds to customers in the podcast as well.

Tune in for an in-depth discussion into the very bleeding edge of data innovation with a company at the forefront of it.

Enjoy the show!

General info about the Data School

Application process and deadline for the next 3 intakes: https://www.thedataschool.com.au/apply/

Learn more about MIP

Thank you to you our sponsor, Talent Insights Group!

Join us for one of our upcoming events: https://www.datafuturology.com/events

Join our Slack Community: https://join.slack.com/t/datafuturologycircle/shared_invite/zt-z19cq4eq-ET6O49o2uySgvQWjM6a5ng

“Like any tech, you’ve got to go through that change management philosophy where people get trained up, they’re expecting it, especially the newer and the more that you break the norm, the more resistance you get, right, so you’ve still got to deal with that, that’s never gonna go away, regardless of what you take you’re using.”

— Peter Kokinakos, COO of MIP

WHAT WE DISCUSSED

00:00 Introduction
05:29 What are graph databases? What are some potentials there? Why is it important?
10:26 How do you recommend that people get into using graph databases, or get started in the space?
11:52 What about the applications or use cases, where should people look for when starting in the space?
23:33 Are there any particular challenges when it comes to graph databases or graph-type applications that are different from normal?
26:19 What are your thoughts and perspectives on the challenges?
33:49 Why or when did you guys decide, to bring that school over to start this? How did that point come about?

EPISODE HIGHLIGHTS

• MIP is the kind of company that for 30 years has been able to tap into some good talent. And we tended to use very experienced people in the data warehousing space. And we still continue to do that.

• So as we're trying to find people, it's tough at the moment, it really is tough to get good people and people with good experience. And what we did, we looked around we thought, You know what, there's a lot of people out there that actually have good analytic talents, but they're not certified. They're not educated with the right initials from about a university that says, here's the right initials to be an analyst and all those good things, yet they're really, really good with logic. They're good at thinking through different angles. And being an analyst. You know, that's what primarily you have to have this natural curiosity to say why, and how are these related and why they related? And how does that look?

• We go through that process and find these people is just an amazing, amazing pipeline of talent that we have on Earth. We've got people with PhDs in zoology, I've got psychology students, psychiatry, students, I've got medical students, I've got engineers, lots and lots of engineers.

• And you know what, they can earn a lot more money as analysis data scientists. And so we've got, I think there's now 150. Through the system in Australia, there are about 120, still in 30. Of the alumni globally, there are close to 500 people, the data school started in London, and opened up in Hamburg in New York recently. So it's got a connection around the world.

•  We teach them all about things like data governance, project leadership, and time management, so we give them all the soft skills that they need to be a great analyst. So they come out there and they're changing the world, they're actually making some big, big strides not just in Australia, of course, around the world with the skills that they bring to our clients.

• I think the Tertiary system still is working pretty well, I think the challenge for us is that timeframe, right, waiting 3, 4, 5 years for someone to complete the mark the bachelor, some of them doing honours, then masters, that kind of pipeline in this kind of marketplace just didn't work for us.

• So we just wanted that immediacy. And the only way we could do that was by controlling ourselves. So that's why we jumped into creating the data school down under which is if anyone who's watching this is kind of thinking I'm learning about analytics, I really want to get a job. But I don't have the qualifications, but I don't have the experience, come to us, and we'll go in a different way to get into the world of analytics.

• If you're thinking about putting together a graph, database application or a system you're in your premises, think about starting with something brand new, don't try and get an old app that's kind of struggling or isn't really working for you. Because sometimes there's too much baggage around that there's too much view from senior leadership as well.

• So you want to try and start with something fresh start with something, again, that the old data science thing about, yeah, making sure you can define the problem effectively is the same concept that applies here, don't just try and throw some data in and hope to find some stuff, you got to have some kind of idea of what you're trying to get out of this.

• Now I've got a really different kind of way to start thinking about my data and the relationships that exist within my data because that's the power of what we're looking for here is how the things connect and how they are related. And that's really what the graph does for us.

• Making sure you can define the problem effectively is the same concept that applies here, don't just try and throw some data in and hope to find some stuff, you got to have some kind of idea of what you're trying to get out of this.

• And I think a lot of the data lakes are forming a very, very heavy, thick layer of ice on top of them. So the water gets in or the data gets in, but it's not coming out easily. And so this is a great opportunity to use graphs to kind of navigate through some of the stuff because one of the advantages of graphs coming in a little bit later in the process in terms of the database technology structures, they've already solved the problem of scalability and size, and all those good things.

• We kind of think sometimes technology is going to be the silver bullet that will change the world. Sometimes it just makes things faster. So we know, hey, there's no value with this. Let's move on.

• But again, like any tech, you've got to go through that change management philosophy where people get trained up, they're expecting it, especially the newer and the more that you break the norm, the more resistance you get, right, so you've still got to deal with that. That’s never going to go away, regardless of what you take you're using.

• Graph is now coming to fruition in terms of bringing value to clients. And not just being theoretical, actually, being a real product.

---

At Data Futurology, we are always working to bring you use cases, new approaches and everything related to the most relevant topics in data science to help you get the most value out of these technologies! Check out our upcoming events for more amazing content.