This transcript is edited from the TigerGraph Connections podcast published August 30, 2022, with Max Latey, CEO and Founder of Pinboard Consulting.
Corey Tomlinson: So before we get into your current work, you were part of a winning team in the Graph for All Million Dollar Challenge earlier this year. Your UN Data Project creates a better way to search and visualize the data published by the United Nations. Can you tell me more about how that project came into being and what your vision is for it in the future?
Max Latey: Way back in the day, during my first few modules on my Master’s in data analytics, I had to pick a dataset to do some work on, you know, whether it be visualization or understand the data processor. And, you know, we had quite free rein from our lecturer and supervisor on this one. And I chose to use the UN data, it was this amazing repository, fascinating data, all kinds of information about economics, health, pollution, tourism, all kinds of stuff. And I thought, well, you know, okay, what a great dataset to work on.
I started working on it and found well what a bad dataset to work on, it really was extremely challenging as a dataset! Still fascinating. And the particular data that I looked at was maternal health. And particularly trying to look at maternal health, by country, by income. At the time, I didn’t really know much about graph; I’d never really used graph. And so I tend to treat it as flat data, you know, as you would expect – tables, linking tables, doing joins, trying to do interpolation, and just found it so hard to work with, but got to the end in my report, and I didn’t really think anything more of it.
Fast forward a couple of years later, and I fell in love with graph theory, fell in love with graph technology, and launched Pinboard, to really try and bring the power of graph to companies all around the world. And the hackathon came up. And in the back of my head, I thought, ah, you know, what, now that myself and my team really know the power of graph, this is a great opportunity to revisit what was a very painful process and make it better. Not just make it better for me, but make it better for everybody. And so that’s what we did. We looked at the different categories in the hackathon. And there were some, you know, really interesting stuff. But we took a vote as a team and went, let’s look and see if there’s a way to get this amazing data resource into some kind of graph technology, where it can become easier to use, easier to analyze – smooth to pull out, smooth to operate with, in a proper graph-native way.
So the hackathon was all about pulling out that data, loading it in developing the schema, figuring out what people might want to do with the data, and making it available that way. We found someone to do a nice UI. We built filters; we built a really nice kind of scatterplot, which would allow people to pick metrics within the data. It might be the ones I looked at originally, which were health and economics, or it could be tourism and carbon dioxide, or it could be, you know, crop yields and nitrogen pollution. This was really what we wanted to try and do with this application – give people analytical power and analytical tools to be able to get information out of this incredible dataset, but without having to go through what I went through originally, which was dealing with these flat datasets and having to join them yourself. Get them into a graph, get them into TigerGraph, and suddenly, the relationships are just there. You can query them at your will.
So yeah, it was a great time. To really make that doable, we came to the end of the hackathon, we had an end-to-end solution, a product and an app that people could use and query with some data in it. It could have used some more refinement, you know, hackathons are time-limited by their nature. And we haven’t had a chance yet to go back to it. But there’s always this plan within Pinboard that we will open the box again when we’ve got a little bit more capacity, some time, you know, we’re able to dedicate some resources to it, that we will make the app from a hackathon ready, rough and ready kind of tape and string version to a nice, industrialized product version where we really can fulfill that promise and get the data into graph, make it available to journalists, economists, researchers, students, in a much, much easier way.
Preferably with some of these tools, which we’d conceived of during the hackathon, like the automatic scatter plotting, maybe time series analytics, there are things which you can just do out of the box, allow people to pick metrics, allow people to pick time ranges, countries, and get this data out there and turn it into information or insight.
Corey: You’ve mentioned your team a few times. How many people were on the team for the project? Was it the whole of Pinboard Consulting? Or was it a select group?
Max: It was all of Pinboard and one other person. So technically, we were a team rather than a corporation, which was nice. That was something that I wanted to do was really work with the community. So we used the Devpost page, found people looking for a team, reached out, and we found a really good UI developer, as it happens in a different country. And yeah, we just worked really well together as a team.
So Pinboard, we’re mostly doing the data consumption, the schema design, and looking at the analytics algorithms, and then our teammate built the UI, the interfaces and helped with suggestions on the plotting. We built a really nice global country selector to allow the country filters for returning the data. It was great working as part of a team, you know, feeling like something wider than just Pinboard.
Corey: So you mentioned that the UN data project actually came from a paper you wrote earlier during your master’s coursework. Now, as I understand, you’re still working on that, currently finishing up your dissertation on human social behavior? How is that going?
Max: It’s nearly there. The journey has been so interesting. It’s an area of graph theory, and an area of data analytics, which, unfortunately, is not that widely used in the industry. So although it’s been intellectually incredibly interesting, fascinating, and fun, it’s not been the most industry-applicable subject I could have chosen, but that’s good. You know, it’s given me something outside of work, but still in the graph space to look at and think about and, who knows, one day temporal graph analytics may become part of what we need to do. But for now, it’s real cutting-edge stuff. I mean, other than maybe some of the giants in the global graph space like Instagram, Twitter, LinkedIn, maybe using some of these temporal techniques, but in the wider industry, it’s just unheard of. So for me, it’s just been fascinating to deal with these cutting-edge maths, cutting-edge technology, and do proper scientific investigation. It’s been a great couple of years.
Corey: You mentioned that was an area of graph that’s not used when you were going about deciding on that project in your dissertation. Did you realize that it was graph-related? Or was that something that was just kind of an ‘a-ha’ moment during the process?
Max: I knew I wanted to do graph. So when I was picking a supervisor and picking a topic I knew I wanted to do graph. Did I know I wanted to do temporal graphs? No. Did I know I wanted to do human social interaction graphs? No. That all came out of some great discussions with my supervisor. I kind of went in with a love of random graph theory, which is where you try to build graphs using nothing but random parameters, which resemble something real. So whether it resembles a social network, a transportation network, a genetic network, a chemical composition network – whatever it may be, if you can generate a random graph, which, without any kind of other identifiers looks indistinguishable from a real graph, then you found something, right. You found something in these parameters, which is universal, which tells you something about these real networks.
And so I knew I wanted to do random graph; I got allocated an amazing supervisor, Ginestra Bianconi, who’s a pretty big deal in the graph space, particularly in these temporal networks. And yeah, we sat down and chatted, we looked through the literature, whatever was current at the moment, so she showed me some great papers that I could use as kickoff points. We narrowed it down to a great paper from 2018 on a random graph model to represent human social interactions.
Corey: Million Dollar Challenge. Master’s degree. And you actually have your job as the CEO of Pinboard Consulting. There are only 24 hours in the day. I was going to ask if you slept ever, but I’m not going to ask that. Can you talk a little bit about your day job as the CEO of Pinboard Consulting?
Max: Graph can solve problems that no other technology can. Companies are now starting to work that out. A lot of companies in a lot of industries have tried to do analysis on their data using flat databases, or using key-value pair databases, like Mongo or Cosmos, and failed. They know what they want to ask. Their solutions can’t answer those questions. But they’re finding graph, and in particular, the only truly scalable graph, which is TigerGraph, and that’s really where Pinboard comes in, is helping companies who know what they want to do, but can’t do it themselves.
You know, you can’t go out to the market and hire ten graph technologists, there aren’t ten to hire, right? If you’re lucky, you’d find one. If you want to do an implementation project, where you want to put in TigerGraph, if you want to plug it into your systems, you want to plug it into your machine learning platforms, you want to plug it into your BI, you want to plug it into any of your other analytics, you want to do these investigations. It’s tough, right? And so that’s where they come to Pinboard and say, “Hey, can you guys help us?” We want to put in this amazing new system. We’ve got these great questions we want to be able to answer, we know graph can help us answer these questions. We just can’t get it in. Can you help us? And so yeah, that’s what Pinboard does.
In terms of where I fit in, as a CEO, I tend to wear a lot of hats. We’re a small company. So one day it could be project management, one day an education talk, you know, extolling the benefits of graph, another day it could be talking to TigerGraph engineers about ideas or feedback from the community. You know, we’re pretty well plugged into the wider network technology space, academia, other industries, other people using graph. My day can be anything from phone calls with stakeholders to sitting on a talk like this, saying how great graph is.
It’s just it’s been a great journey. You know, very few people get to work in their passion place. And I get to do that, you know, I get to come in and work with graph technology, graph theory, graph technologists, people who are just as keen and enthusiastic as I am. And this is why I changed career. This is why I shifted from QA and test management in technology, which was my background, into the graph space. It was to do this. And now I get to do it. I’m a very lucky guy.