The Data Stack Show - The PRQL: Graph as a Utility

Episode Date: November 14, 2022

In this bonus episode, Eric and Kostas preview their upcoming conversation with Ryan Wright of thatDot.

Welcome to the Data Sack Show prequel. We just recorded a really interesting show with Ryan Wright from a company called BatDot. They are the commercialization of an open source technology called Quine. And what an interesting conversation, Kostas. We talked a lot about graph as a concept, but then also graph databases. What was your takeaway? Because we haven't talked about graph on the show a ton and it was a really helpful conversation,

at least for me, but what was your big takeaway? Yeah, I mean, graph databases and graphs in general is like one of these things that they are not that much discussed today. I mean, there are like some technologies out there that are like, okay, they're doing pretty well, like people are using them, but it's almost felt like there are some kind

of like niche. Yeah, solution, I agree. But what I really enjoyed, like with the conversation that we had is that it's a very refreshing point of view in terms of what and how graphs can be used. And they don't necessarily have to be, let's say, these monolithic database where everything is only a like a node and as a graph. But we can take like a different take on it and use it like more as like a processing paradigm and a query paradigm like for specific problems and do that like on top of existing systems that they don't look like necessarily graph-based systems. Right.

And that's like what I I found extremely, extremely interesting about weighing the full conversation that we had. And yeah, I really enjoyed that. And I'm really looking forward to more examples of this approach, to be honest. Yeah, I agree. I think the teasers here that I think are really interesting are that the Quine team's take on graph is less of a self-contained unit that does a very specific type of thing really well and almost looking at it as a utility. And one of my favorite examples on the show was actually using Quine as like a pre-compute engine, you know, to save costs downstream by leveraging the graph to do like some complex

initial compute in a way more cost-effective way, right? And at that point, it's almost, you know, really is just sort of a transformation layer in a pipeline. You know, if you think about like the basic data flow, but graph happens to be like super well suited for that like use case. But to your point, no one really talks about graph like that. So all that to say, definitely check out this episode. It's one of the most compelling, you know, sort of conversations I think I've heard on new ways of thinking about using graphs. And subscribe if you haven't, of course, and we'll catch you on the next one.