Making Sense of Unstructured Data: Expert System Interview

Making Sense of Unstructured Data: Expert System Interview


Making Sense of Unstructured Data: Expert System Interview Transcript

Come learn about the fascination intersection between Semantic Intelligence and the Oil and Gas industry. To learn more about Expert System click here.

Mark: Hey, folks let’s learn something new about the oil and gas industry.

All right. Today, we set it up a little bit different. Bryan and I just cannot figure out how to get in the same town at the same time, so we decided to do this interview via Skype. So, good morning, Bryan.

Bryan: Good morning, Mark. Thank you for having me today.

Mark: Oh, absolutely. So, where are you geographically?

Bryan: I am geographically located in Chicago, Illinois.

Mark: So, how’s the weather today in Chicago?

Bryan: You know I’m looking outside and if might be able to tell by the reflection behind me the picture, then it’s probably 75 degrees and beautiful and sunny.

Mark: Yeah. I’m so jealous. It’s 97 degrees here, 90% humidity already, it’s like ahh, the summer has already hit.

Bryan: Well, that’s why we’re doing it via Skype. I don’t want to be in Houston right now.

Mark: Yeah, touché on that one. So, Bryan works for this really interesting company called Expert System and they do something called – help me with this Bryan, is this semantic intelligence?

Bryan: Yes, semantic analysis of content and what it comes down to is really being able to look at structured and non-structured content, but more specifically unstructured and words of context.

Mark: So, let’s talk about that a little bit. So, what does that actually mean for a company?

Bryan: What it means for company is most organizations have either keyword they just deployed or keyword technology or statistics-based. And the challenge they run into a word ambiguity. And to give you one example, take the word stock. If I said the word – a sentence to you, I have – I bought ten thousand shares of stock in Apple Computer. Like you understand what I mean by stock and Apple, but if I change that and I say I have ten thousand apples and stock, I’ve just changed context. So, apples and stock take on new meaning which is incredibly important to the oil and gas industry. It’s understanding that ambiguity.

Mark: Yeah, absolutely. So, you know, there is terminology in this industry that exists nowhere else. And so, you’re telling me that your system can actually go in and look at the unstructured data in the oil and gas company and make sense of all these stuff so that when people need to find it, they can find it?

Bryan: Correct. What we really focus on is adding dynamic metadata to unstructured content and therefore it’s structured. And now we can also start co-mingling the unstructured with the structured.

Mark:  Yes. So, not to get too far over our audience’s head, so basically the structured data is data that you have a descriptor about which is what you’re calling metadata. And unstructured data, you don’t have that description, is that close?

Bryan: Correct. The unstructured data usually has a few pieces of metadata in there, you know like a date or word or document type things like that, but there’s not a whole lot of information around the unstructured texts itself.

Mark:  So, if an oil and gas company had, you know, twenty or thirty years of unstructured data in paper records and they scan them and use optical character recognition so they actually have this data, your system can go in and my understanding is actually read it almost like a person would and understand where it fits?

Bryan:  Yes, it could read the combination of linguistics and semantics. And that is example we used is perfect because one of the big deployments in the oil and gas industry this year is about analysis. So, these organizations have network shared drives spread out around the world and there’s a lot of valuable content on there, but the oil and gas industry hasn’t been leveraging it to its whole potential. So, that’s one of the deployments is analyzing network shared drives.

Mark: Man, that is huge. So, right now this low crude price is operational efficiency especially for upstream is all of a sudden important. You’re building for your people to find what they need as quickly as possible affects the bottom line and you actually – you actually impact that in a big way by making sense of their unstructured data.

Bryan: Right. This is about making a content viable and reusable, but also if you have the ability to understand the contextual relevance of the terminology. One deployment was through the health safety and environmental taxonomy that an organization built and what they discovered was they’re trying to determine what type of accidents were occurring at well sites. And what they discover were there are a lot – there’s a lot of information about burns, but what they quickly figured out is there’s a difference in burns; some are hot burns, some are cold burns, and having the ability to analyze the content correctly enables them to look at the information more accurately

Mark: Man, that is huge because you know as well as I do that HS needs a major driver in this industry. Man, this is like rocket scientist.

So, Bryan if somebody want to learn a little bit more about what you all do, where should they go?

Bryan: They should go to

Mark: Yeah. And folks, we’ll put a link in the show notes, so you don’t have to be taking notes.

Bryan, thank you so much for your time today.

Bryan: I appreciate it, Mark. Thanks for calling.

Mark: Yeah. Folks, I hope this helped, we will see you next time.