Episode Transcript
Derek Robinson 00:00:01 Welcome to Data Points, a podcast by InterSystems Learning Services. Make sure to subscribe to the podcast on your favorite podcast app. Links can be found at datapoints.intersystems.com. I'm Derek Robinson, and on today's episode, I'll chat with Carmen Logue and Mary Ann Fusi about the latest in InterSystems IRIS Adaptive Analytics.
Derek Robinson 00:00:34 Welcome to the Data Points Podcast by InterSystems Learning Services. On today's episode, I'm joined by Carmen Logue and Mary Ann Fusi, who are both in the data analytics space here at InterSystems. Specifically, we'll be talking about InterSystems IRIS Adaptive Analytics. Now, you may recall that Carmen joined the podcast last year in Episode 17 as part of a roundtable discussion we had about the various analytics and AI technologies within the InterSystems stack. Thomas Dyar and Benjamin De Boe were also on that podcast episode, talking about various machine learning and data analytics topics. A lot has changed, though, since that episode, and improved, frankly, since that time. So, Carmen and Mary Ann are here today to share the newest features, customer use cases, and more. This conversation rolls into a larger one, which is rippling through several industries right now, and that is around the concept of an enterprise data fabric—or what InterSystems calls a smart data fabric. Part of this conversation is around shifting the mindset to have data from various sources in your organization accessible at the time of need, organized enough to aid business decisions, retrievable without the latency of batch processing or manual consolidation. We have linked in this episode description a recent talk that Carmen was involved in, as well as Joe Lichtenberg from InterSystems, about this very concept of a smart data fabric. That presentation was from AtScale's Semantic Layer Summit in September. So without further ado, let's hear from Carmen and Mary Ann on Adaptive Analytics.
Derek Robinson 00:02:06 All right, Carmen and Mary Ann, thank you both so much for joining us to talk a little bit about Adaptive Analytics today. So Carmen, I'm gonna start with you. Many of our listeners, from season one over a year ago, we did a podcast on the analytics landscape at InterSystems, different AI technologies, things like that, with you and a couple of others in product management. So before we dive in, let's, can you just give us a little reminder and kind of set the scene for what Adaptive Analytics is, so that we can have this kind of follow up conversation about it?
Carmen Logue 00:02:34 Sure. I'd be happy to. Yeah. I think at the time we did that podcast, Adaptive Analytics had just come out, you know, so it was, so we were excited about getting started with that. And Adaptive Analytics is really a layer on top of InterSystems IRIS and IRIS for Health that provides some standard definitions. We call it a semantic layer, and it provides kind of standard naming and definitions for mostly business intelligence users. And so business intelligence users are using Microsoft Excel or Tableau or whatever tool they want, and it gives them kind of a common interface to the data that's in IRIS, and allows them so that, you know, for example, somebody who's doing a calculation in finance on Excel will get the same answer as somebody working in Tableau in marketing. So it gives, uh, that common kind of layer as well as doing some performance enhancements to enable faster queries as well. So it's creating aggregated data for common queries that people are gonna ask over and over. So it's really about both performance and consistency.
Derek Robinson 00:03:44 Right. Nice, nice. And, and so you mentioned semantic layer was one of the things in that conversation as far as describing what Adaptive Analytics is. And so the Semantic Layer Summit is actually something that recently took place with our partner AtScale. So can we talk a little bit about that?
Mary Ann Fusi 00:03:58 Yeah, absolutely. AtScale, who is our partner, who provides the semantic layer technology, held the first annual Semantic Layer Summit on September 28th to, they brought together more than 30 thought leaders and industry experts to talk about the importance of using a semantic layer with data and analytics. We were really happy to be part of that, and Carmen and our colleague Joe Lichtenberg held a session to describe how organizations can unlock the power of the data that they already have by utilizing a smart data fabric architecture tightly coupled with a universal semantic layer.
Derek Robinson 00:04:47 Nice. That's great. And yeah, and so some of that, smart data fabric, you know, we talked about in the intro of the podcast a little bit, and there's actually some related materials we'll talk about at the end of the episode, that get into some more of that from Joe and both of you, with that conference. So, looking at kind of Adaptive Analytics now, right from, you know, after talking about it a year ago, what are some of the new capabilities that have been introduced into this product? You know, as the InterSystems analytics and AI landscape has evolved?
Carmen Logue 00:05:16 Yeah, so Adaptive Analytics became available with InterSystems IRIS and IRIS for Health in release 2021.1, initially. And then in the next release, uh, we added a couple of features. One of them was to enable InterSystems Reports as another BI tool. So just like somebody can use Tableau or Microsoft Power BI, they can also use InterSystems Reports, and they'll get the same set of definitions and the same set of, uh, calculated measures, all that stuff, to be available in InterSystems Reports too. So it really kind of ties together some of our analytics offers. Uh, the other thing in that same vein is that we allowed, we made the ability to export from an InterSystems IRIS BI cube into Adaptive Analytics as a virtual cube. So it's also kind of tying together some of our analytics offers and, you know, allows people really an opportunity to get started with Adaptive Analytics, if they've got some InterSystems IRIS BI cubes that they wanna try out there. So it's another kind of integration capability. In addition to that, we've also done a lot of work around performance tuning around scalability as we've, you know, deployed in a number of locations, we've been able to learn and get better and find new ways to integrate with AtScale, and have improved both how we tune IRIS as well as tuning AtScale and Adaptive Analytics. And the last thing is, we're just getting started with something called AI Link, and I know we're gonna talk about that a little bit more. We're pretty excited about that, too.
Derek Robinson 00:06:57 Yeah, Nice, nice. And so, you know, I think that's one of the cool things about, so I talked to you offline about how we did this podcast, we launched it, and we kind of covered all the basic technology areas, right, at InterSystems…it's almost like a "what is" level episode with a lot of these. This is one of the first ones now we can follow up and take a little closer look at okay like, how is some of this being used? Right? And so I think, one of the fun things to do, and so just, I've mentioned this past episode a few times, listeners can go back to Episode 17; that's the reference point for the analytics and AI, uh, landscape discussion we had. But moving into some of the cool use cases that maybe we've seen some of our customers and partners use, and understanding we can't name them necessarily, but what are some of the coolest things that we've seen implemented now with Adaptive Analytics?
Mary Ann Fusi 00:07:42 Yeah, there have been quite a few, actually. One example is a longtime InterSystems customer who is using Adaptive Analytics with IRIS for Health to provide hospitals with insight into the drugs that are most often used within their hospital, and how frequently those, the recommended dosages are being overridden. It's part of a new business initiative that will enable the hospital to unlock the data, unlock the power of the data, so they can provide actionable insights.
Carmen Logue 00:08:24 Yeah, and I think this one is, this is a really interesting one, because they're a long time InterSystems customer, and so, you know, IRIS has kind of been the backbone for some of their other solutions, and this is a new offer that they're gonna be able to provide to hospitals, like Mary Ann said, that, provides, you know, analytic insight into what they're seeing, you know, and dosage information, as well as opportunities for additional training at the hospital, so they can really take control of the data.
Mary Ann Fusi 00:08:56 And then ultimately provide better patient care.
Carmen Logue 00:08:59 Yeah, exactly. Exactly. Uh, it's, you know, another one I think this reminds me of is, also has been a long time, started as a Caché customer and then moved to IRIS several years ago. And, you know, they've been using us for kind of all their, you know, the operational and transactional data and, you know, when we were in the process of doing a proof of concept with them, you know, I asked the Sales team, you know, what does this, what does this give them? What additional, information this give them? And they said, well, this is really opening up visibility to data that they haven't been able to see. And that really struck me because, you know, they have, they've been using some of our other products for some insight, so they do see some things, but they're, you know, they're focusing that on, you know, very specific areas, but they had a whole other team in their organization doing analytics and, you know, kind of bringing in this proof of concept for Adaptive Analytics allowed them to expand into other parts of the organization that had been running alongside IRIS for quite some time, but not, you know, were not really using the data that was available. So that was, that's another kind of interesting, uh, use case, was just providing insight into that data that's been there all this time, but they haven't really been able to unlock.
Derek Robinson 00:10:19 Right. Yeah. And that seems to play into what you mentioned earlier with kind of that smart data fabric and, you know...maybe just there's a lot right in front of you that actually is there in your organization that you can take better advantage of, right?
Mary Ann Fusi 00:10:30 Exactly.
Derek Robinson 00:10:30 So kind of, uh, going back to one of your earlier points, in the last episode that we've referenced a few times, people may have heard Tom Dyar, who's one of your colleagues in product management, talking about machine learning and IntegratedML during that AI discussion. So you mentioned AI Link, which is one of the new capabilities that you mentioned, and just, you know, at a detailed level, we're gonna bring Tom in for that conversation at a different time. But with an Early Access Program for that running, tell us a little bit about what that is, and why people might be interested in that.
Carmen Logue 00:11:02 Yeah, yeah. So, AI Link is an additional feature with Adaptive Analytics that really prepares and allows access to the data for the data scientists through something like a Jupyter notebook. So just like Adaptive Analytics provides that common semantic layer to all BI users, it also can provide that same information to people, you know, to the people who are doing machine learning work. And the advantage there is that they spend less time curating the data and managing the data and manipulating it, and calculating it. So that's one advantage. But the second advantage is they can also write back predictions to InterSystems IRIS, and make it visible in that same semantic layer. So if I'm working in Tableau, I can actually have visibility to some of those predictions, and I can include those right in my dashboard. You know, so it's pretty cool, actually. So, you know, I have immediate access to those predictions. If I notice something as a BI analyst, I can let the machine learning team know, and they can adjust their calculation and then update, you know, update those predictions based on maybe some new information. So it really draws together the AI and BI teams, in a way that I think is, is gonna be really interesting.
Derek Robinson 00:12:24 Yeah. I think in some of those earlier discussions about IntegratedML and some of the learning content we have available, a lot of it is very, you know, sort of limited to SQL query, right? Like, you're basically getting those predictions as part of a query you're running, you kind of see the results, but then, you know, you go to the next page and those results are gone unless you're gonna run that query again. Right? So I think that kind of tying that back into your whole landscape is a really, really cool capability.
Carmen Logue 00:12:49 Yeah. It is. Actually, you mentioned the EAP. We should call that out here too. We're, you know, at the start of an Early Access Program, so looking for more people to participate in that. We're doing it as a hosted EAP. So what that means is we're trying to kinda lower the barrier, or the barrier to entry, for participation by hosting both Adaptive Analytics, IRIS, um, and so what you can do is bring your own Jupyter notebook, get access to, and actually we've loaded up some, uh, some standard data too, so that you can actually use that existing data set. If people are interested in loading some of their own data, that's also a possibility, but that way, you know, there's no installation that you have to do. So we're really looking for, we're excited about getting some people started as part of that hosted EAP.
Derek Robinson 00:13:40 Nice. And, just out of, you know, my own curiosity, how is that hosted? Is that hosted on like AWS?
Carmen Logue 00:13:45 AWS, yeah.
Derek Robinson 00:13:45 Nice. Is that kind of tying in with our kind of cloud service offerings that you're seeing? Cause we're, that's one of the things that kind of InterSystems and Learning Services have been talking about quite a bit with these different cloud service offerings through the InterSystems cloud portal. Is that kind of the same environment that this EAP is?
Carmen Logue 00:14:00 Yeah, yeah. It's like that. So, yeah. I'm hoping a lot of people sign up and check it out.
Derek Robinson 00:14:05 Yeah, absolutely. So the link for that is gonna be in the episode description. Uh, you can follow up with that, but just kind of pulling back in general to, you know, kind of wrapping up the conversation on Adaptive Analytics, aside from this EAP, how can people learn more, and what should people go seek out kind of to stay up to date with new things that come out for Adaptive Analytics?
Mary Ann Fusi 00:14:22 We have a lot of information on our website, and I would invite all the listeners to check out the resource page of our website in particular, and just search on Adaptive Analytics. You'll find the recording of the session that Carmen and Joe did for the Semantic Layer Summit, as well as a new e-book that describes how you can unlock the value of your data, as well as a bunch of other resources. And then there's a bunch obviously about the smart data fabric as well, that kind of ties in nicely with Adaptive Analytics. And as was already mentioned, the EAP link at the bottom of the episode description.
Derek Robinson 00:15:06 Cool. Awesome. Well, Carmen, Mary Ann, thank you guys so much for joining us, and we'll talk to you soon.
Mary Ann Fusi 00:15:11 Thanks.
Derek Robinson 00:15:14 Thanks again to Carmen and Mary Ann for joining us for that conversation. As you can hear there, there's a whole bunch of exciting stuff going on within the data science and analytics spaces at InterSystems. Make sure to check out the links in the podcast description for the Semantic Layer Summit talk I mentioned earlier, as well as the Early Access Program that Carmen mentioned. That'll do it for our Adaptive Analytics episode. We'll see you next time on Data Points.
