Drill to Detail - Drill to Detail Ep.43 'Oracle Analytics, Data Visualization Desktop 4.0 and The Art of Product Management' with Special Guest Mike Durran

Episode Date: November 30, 2017

Mark is joined by Mike Durran from the Oracle Analytics Product Management team in this UKOUG Tech’17 special to talk about his route into product management via the Oracle Discoverer BI tool, Oracle’s latest product in this space Oracle Data Visualization Desktop 4 and its new features, and Mike’s upcoming sessions at the UK Oracle User Group’s Tech’17 event next week in Birmingham, UK.

So welcome to Drills of Detail and I'm your host Mark Rimmel. I'm joined today by Mike Durran who's a product manager in the analytics product group at Oracle. He's also an old friend I've known for about 10-15 years now. Thanks Mark for eventually inviting me onto the podcast. It's good to be here. Well we ran out of interesting guests so you were available so we thought I thought you'd get you on and there we go. So how long have I known you now? Must be about since, must be 10-15 years. Yeah it's getting on to that. I um i was looking through my email archives when you invited me on the podcast to see if i had any dirt on you um i think the uh and this is i believe going out around the time of uk og and i think the first time i went to uk

og in birmingham we actually did a co-present on some topic i think you were presenting about the obi ebi server and we were talking about integration of bi Publisher with Discoverer. So yeah, and that was 2000 and something, small number. Blimey, and then going back in the years, I mean, you've been at Oracle for a while now, haven't you? I mean, tell us how you got into Oracle, what was your route into there when you joined Oracle? So my background was in computational chemistry. And what that means really is that one day, I was, you know, building a program to rotate molecules on the screen,

3d kind of graphics, then the next day, I'd be in the lab, you know, doing organic synthesis, or building a database to store results results so i i you know i always had that kind of interest in and computing and you know essentially using it to solve problems um and also one of the the key things i was kind of quite interested in was things like simulation so i went on to do research in simulating carbohydrates and essentially trying to find trends within the data. So I spent a lot of time wondering what it would be like to be an electron around an oxygen atom with a magnetic field running through you. Spent three years thinking about that.

So Oracle was an attractive company to work for from that perspective because of those uh you know similar kind of uh topics really in terms of reporting analysis but you know i'm i'm coming straight out university and i'm pretty green when it comes to uh you know the IT industry and technology. So I got accepted onto the graduate program at Oracle. And that was long enough ago that you would actually spend a week or more than a week sometimes going through pretty much the entire product set. So they gave you this bootcamp

and you'd learn all about SQL. You'd learn about the database and DBA skills pl sql reports and forms um and then you know the challenge then as you when you come in as a graduate is to actually as a graduate consultant is to actually get out onto a project so catch 22 you can't get out onto a project until you've got experience and you can't get experience until you go out into a a project. So I did, fortunately, I did a few more training courses in the OLAP technology and also in Darwin. Do you remember the Darwin data mining tool? Yeah, exactly.

Thinking machines. So yeah, so I spent a lot of time learning and I even did a C++ course so you know in my background i'd done c programming and that it got to a point where i had a phone call from from my team leader i happened to be on a training course uh playing five-a-side football in the evening and i got a phone call and it said okay we've got an internal project for you and it's working with the the HR team in Bracknell and there's the there's a new tool that's just come out called Discoverer and they want somebody to go along and build a system for them using this this new tool and and since that day, really, as a consultant, I didn't have one day on the bench.

So that was, I guess, my route into Oracle. And, you know, my first experiences really of working with that discoverer product. Okay. Okay. So you told me once, I think it was over a beer at an event. Actually, the reason you ended up at Oracle, I'm not sure if you were joking at the time, was you actually mistook him for the name of that teletext company. I'm not sure whether you were kind of winding me up at the time,

but what was the story there? Yeah, no, I absolutely was winding you up. I think the joke really is that, you know, back then when you said that you worked at Oracle, people did think that you worked for a teletext company and that you were typing the news that they saw on their TV if they pressed the red button or something. Yeah, exactly. Or maybe the Oracle Shopping Center in Reading.

Well, that as well. And just to make it worse, it's in Reading, which is where the Oracle UK headquarters is. Exactly. Exactly. So, Mike, the reason I asked you to come on the show was to talk about Oracle Data Visualization Desktop and the very recent version 4 release. And we're going to spend most of the time talking about that. But you talked about Discoverer there, and I think there's kind of parallels between Discoverer and DVD

and their desktop tools and their they're kind of uh you know for analytics and so on but what so discoverer was you fondly remembered wasn't it by people and it was certainly innovative in certain ways i mean what what were some of the things that you were kind of involved in there and and what the things you think were good about discoverer that were maybe perennial good features and things that that still be relevant now so discoverer that were maybe perennial good features and things that that would still be relevant now so discoverer was a great way for me coming into oracle like i say as a graduate and it was a great way of getting to know a system as well as building up skills talking to end users to try and understand what they wanted to get out of their data. So, you know, I don't, obviously, I don't need to tell you about being a consultant.

However, one of the key things that I think is, you know, a key skill about being a consultant, really, is being able to translate an end user requirements into a technical um you know implementation um and i'm i don't know whether you had this experience as well but it was around the time when um discoverer you know somebody would buy discover and you'd go in for a few days to um tell them how to use it really and you, you know, a short couple of days engagement. And they would give you a line report, you know, those where you had alternating green and white lines printed out in a dot matrix and say, can you do this, please, in Discoverer? And you'd actually sort of explain to them, well, you know, it's actually you can do a lot more with Discoverer and you can actually interact with the data, you can drill, you can pivot. And in terms of some of the innovations that Discover brought to that

product set and that type of capability were under the covers. So it was about SQL generation, it was about generating the SQL so that it was effective. It was about being able to create summaries where you could then speed up the response time for the end users. And then when they built a query, it would redirect to the summary so it would come back a lot quicker. It was about being able to predict the time that a query would take so that the end user would know whether to go off and get a cup of coffee or whether they would be able to

set, you know, be able to say to the system, okay, if this query is going to take over, you know, a certain amount of time, then I'd like to schedule it and run it overnight. So those kind of innovations, I guess, in some ways were brought to the product. And it was around the same time that analytic functions just started coming into the Oracle database. And one of the areas that I worked on, which was quite enjoyable, really, was building the sample workbook, video stores, around, you know, how to just basically just a number of examples for using the analytic functions for creating top end, bottom type, bottom end type reports, moving averages, you know, those sorts of analytic capabilities, which

before the analytic functions were quite tricky to do within SQL. But then when the analytic functions came into the Oracle database, we very quickly embedded those into Discoverer so that end users could then use that power to get insight into their data, which is kind of, you know, what we've been doing ever since. Okay. Okay. So you worked in consulting, first of all, and then you made the move into product management. I mean, that's actually been something I've been doing recently. I've gone from consulting into being a PM.

And what's interesting is with the product, was that then useful for going into product management? You know, what did you bring to it, do you think, that you wouldn't have had if you'd gone straight into that, say, into that PM role straight from external, for example? Yeah, that's a good question. You know, there's always been a debate, really, about what is the best background to have when you move into product management.

Some people think that pre-sales is a good testing ground for product managers. And, you know, there's others that equally will say consulting. And I think from my perspective, coming into product management from consulting, you're coming with a really good understanding of what the customer is looking for out of the product. And I was probably one of the initial consultants, if you like, for Discoverer. And I spent about four years. I also worked with Sales Analyzer, you know, the OLAP tools from Express, as well as some ETL tools to, you know, build up all kind of, you know, various aspects of the system.

But I guess being one of those kind of initial people working on discoverer you kind of get a bit of a an attachment to the to the tool and i was literally working on a project at some point and one of my colleagues said hey did you know that there's a job going uh for a discoverer product manager in bristol i was sort of living in gilford at the time um time you know had no real intention of moving down towards Bristol but it just felt like one of those sort of destiny jobs really that come along when you think I need to I need to go for that and you know fortunately I got off of the job and moved to Bristol in I think

it was the year 2000 so at the turn of the century I moved to Bristol in I think it was the year 2000 so at the turn of the century I moved to Bristol and sort of working in product management yeah interesting so obviously discoverer was the product then and you were the PM for it and then a few years later you that actually I think you sent me an email you took me off and said you know with the recent acquisition of Siebel and there's a product called Siebel analytics I think you sent me an email, you took me off and said, you know, with the recent acquisition of Siebel, there's a product called Siebel Analytics. I think you said to me, check out that product because I think it's going to be quite important.

And at the time, I remember writing the blog post about it at the time and thinking, this is interesting. This is quite different to Discoverer. You know, it's obviously a different tech stack and it's from Siebel. But, you know, it came along and it became adopted as Oracle's, I suppose, new SQL BI Enterprise Edition. I mean, how did that go down?

Not how did it go down, but what were your thoughts at the time that happened? And how was that technology and team and thinking absorbed into what was now Oracle BI? Well, yeah, that's, I guess, a whole kind of podcast in itself. So essentially, we were working on the next generation of Oracle's analytics tools at the time. In fact, I was working with one of your previous guests, JP, on the administration side, so the metadata building outside.

So anyway, so the acquisition happened. And as you said, the Siebel Analytics tool started becoming more prominent within our product offering. And what was interesting, and one of the reasons why that was the case is the, so Discover was like a client server type architecture. It made direct connections to the database. Siebel Analytics was sort of an N-tier architecture. It was built for the internet kind of, you know, internet world that we were starting to be in at that point in time.

So the architecture kind of really mimicked what we were already building out. So they already had a product that was pre-built so it made a lot of sense to you know move behind that product and you know i think what's interesting really just to cut a long story short is that a lot of the capabilities that we were building out in it was called project armstrong you may remember that name being used so that a lot of that ended up being put into obie 11g eventually anyway so you know none of the work was necessarily wasted it just kind of got repackaged in some respects to obie and then you had a bunch of similar people come in to bring it to be the new pms and and this

sort of thing and one of them actually, Paul Rodwick, he's actually my manager now where I'm working, he actually left Oracle and he's now working I guess into there but it's, what was it like again, what was the culture difference like and what was the, I suppose that process must have been interesting about kind of, you know, I suppose merging the cultures and merging the technology and bringing things forward and so on. Yeah well you know as you say, you know, it's a great bunch of people that came in from Siebel. A number of those people are still around. As you say, I worked for Paul

for a number of years in his team. And a lot of them are now good friends. So. Yeah, yeah. It's been interesting times, isn't it? I mean, so, I recall BI Enterprise Edition, what that became. I mean, we're going to talk about data visualization desktop now, but is the kind of full-scale Oracle BIE, is that still developed?

Is it still going on? Is it still a product that you guys are using at the moment? Oh, absolutely. Yeah, yeah, absolutely. So the on-premise is version or on-prem of OBIEE is yeah absolutely still still out there you know we've still got a large customer base who use that we had a 12213 for it so oh of course yes how could I forget yeah I spent five years of my life writing that book I think there were I think there were about three releases during the actual period of of that writing that book. I think there were about three releases during the actual period of writing that book. And I remember rewriting the section

on metadata development about three times as you changed the, what was it called? The thing about when you have multiple users, that changed about three times in the five years I wrote it. Good memories, you know, good times. But we can talk about DataVis Desktop now in this talk because I'm conscious that kind of, you know, this is where the,

well, to me, it's where the action is happening. Absolutely. So tell us, let's just set the scene first of all. What is Oracle Data Visualization Desktop and where did it come from? And I suppose, what's the purpose of of it and we'll get into the kind of positioning and all that kind of okay so you know as a product manager you know that products are really dependent on you know what is the market needs at that point in time so let's talk about the market and the market context around you know the requirement that has led to this type of analytic tool.

So as we've all seen, the companies and the startups that are successful, so Airbnb, Uber, Lyft, they're all data-driven. So basically disruption within a marketplace is driven by being able to get very quick and timely insights from from your data um and also use that to you know for marketing purposes as well so you know there's a an example that i've seen from spotify where they've used their their understanding of how their customers listen to music for an advertising campaign, where they had a big billboard that said to the user who played Sorry 42 times on Valentine's Day, what did you do? So, you know, there's loads of stories out there about how customers, how

these kind of disruptive companies are able to really make a competitive advantage out of their data. So that's essentially the market context that we're in. And we've got a product offering that we call Oracle Analytics Cloud, OAC and you know that really provides the full stack of analytic capabilities through data preparation, data visualization through to data storytelling. So as the name suggests it's you know a cloud-based product customers can go to our website they can sign up for a trial they can test out it's got you know a lot of capabilities for connecting to you know almost 50 data sources i think at the last count and uh you know i believe

we're going to talk a little bit more about some of the the more you know the new capabilities that have come within that product but the the desktop edition which was released about two weeks ago, I believe, of OAC, effectively DV, Data Visualization Version 4, that introduced quite a lot of new changes to the product. So the first data visualization products, in particular the desktop version, was released probably about 18 months ago. And, you know, the user base that we're looking to satisfy there are people who are looking to do some agile analytics. They're looking to build some visual analytic capabilities from their data. They may want to be, you know, doing these types of analyses offline. So being able to download a tool to your desktop gives you that capability. It also gives you an element of, you know, container privacy. You're doing it simply on your desktop, but also

at the same time, it's kind of fully integrated with, you know, the main kind of product stack. You can share projects with colleagues. You can upload them to the cloud if needed. So that's kind of essentially where the context is around the product and the market background. So the need for an agile analytics tool, which allows you to report against data coming out of an enterprise system and in addition mash up data from other data sources or from files that you've got on your desktop okay so is this I mean if you imagine a member the days of when I was doing IBI consulting and the typical OBI user was a corporate user who was typically I suppose consuming reports that would be built maybe using a BI apps or governed I suppose corporate

maybe kind of finance or kind of ERP type sort of user base. Do you see data visualization desktop as being a different category of user is it a new type of user that is new to the product or Or is it the same ones using, you know, who is the user persona? Are they new or are they existing? Yeah, it's a good question. It's probably a bit of both. So you're going to have customers who,

as you say, they're going to be maybe taking data from a dashboard. And, you know, dashboards are not going away anytime soon. But the ability to be able to take what they're learning from the dashboard and maybe integrate that with some data that, you know, obviously we live in a cloud world these days and people have got other systems that they may have subscribed to at a, you know, line of business level. And

they've got data that they're running their business in, but it's coming from a cloud system. Yet they've got this corporate system and they want to bring the two sources data together. So that's kind of where OAC really shines in making it easy for that type of user to bring data in and then do data preparation on their desktop. They don't need to wait for a process that they would need to go through to actually officially bring online a data source. So that's where the market has gone. So that's where we are offering these sorts of products. A default tool of this category that people would use would be something like Tableau? Why would they use DVD rather than Tableau? Or another tool of that category, not specifically Tableau. Or another tool of that category, you know, not specifically.

Yeah. So, you know, we've got a couple of advantages in the sense that, you know, we believe we're easier to use. So in terms of actually performing an operation, there are less clicks to achieve the same sort of result. And if you're reporting against data from, for example, like Fusion applications or some other Oracle SaaS type system, then we can build native connectors to those types of systems. So there's a couple of reasons why we would go with OAC in those scenarios. Okay. So the other thing that's interesting, I think there's two particular things that are interesting to me with with dvd is first of all the amount of kind of i suppose effort and innovation and and iterations of the product the rate at which the older kind of more uh obi you know the updates to that were kind of fairly kind of big deals and once every kind of like couple

of years or whatever you know there's been releases of dvd very frequently recently and the amount of i suppose new features coming along has been very interesting um but also the fact it is so open compared to other kind of to other data sources and looking at it the other day i was going through using it myself and you've got support for redshift you've got support for google analytics i suppose uh that the oracle's philosophy or is it not i don't know in in terms of data sources. So funny enough, we were chatting about Discover earlier on. The original versions of Discover actually had an element of ODBC connectivity. And then there was a period, as you say, where we focused just on the Oracle

database as a data source. And that was a long, long time ago now. Certainly since probably about 10 years or more, we've actually been open to other data sources. But those types, those data sources that you really need to support have actually increased massively over the last sort of five years. So that's why we now support, as I say, nearly 50 data sources

within DV Desktop and OAC. of five years so that's why we now support you know as i say nearly 50 data sources within dv desktop and and oac um and again you know these other cloud providers you know we we're not going to say well you know we're not going to let you get data out of there we fully want you to get data out of there and we'd like you to do it using our tool so that's essentially where where where we are there so so i mean i connected to bigquery the other day actually so using the j the odbc drivers you can download from google sites i installed those on the desktop and then dvd connected to that and and i'm running against uh against bigquery there which is which is kind of great and that's good so is there in terms of the architecture of dvd is there like is it straight sql access to these sources or do you have like a an intermediate kind of cache

layer or something how does that kind of work yeah so um you've got a bit of both really so you can actually you know pass down um queries to the database which are then cached or you can actually keep the connection alive so start you know keep the keep the data can bring it back so there's basically a couple of options in there for tuning the uh the performance side of things okay okay so let's talk about um just before we get into the new features in in version four so in terms of the packaging and so on of this i mean you mentioned rac you mentioned there's also dv cloud server by dvd on its own or they get it as part of a kind of like a stack or part of a kind of subscription how does it work packaging so simplest way is if you subscribe to OAC Oracle

Analytics Cloud then you'll be able to get access to you know the data visualization desktop version if you're an existing customer so for example if you're using OBI EE on premise, then you would buy the data visualization option. And any kind of licensing questions, I generally refer you to your Oracle contact, because everyone's got a different sort of situation in terms of licensing. And, you know, they'll be able to talk to you about the options that you have. Okay, so basically, it's one of the various options, delivery mechanisms you can use if you have the kind of Oracle analytics stack,

whether you choose to do it on the desktop or in the cloud or whatever, it's part of the same thing. So it's whatever is the, I suppose the delivery mechanism most appropriate for you really. Yeah, what I'd suggest is if you listen to this podcast and you're not aware that Oracle even had this type of desktop tool, go along to OTN, the Oracle Technology Network, you know, do a Google search OTN Oracle DVD. version 4 that we're talking about for Mac or for Windows and install it and

you know just play around with it start getting to getting to know how it works okay okay so let's go on some new features then so I was you know surprised at how much has changed in this in this version 4 and and there's a few things in there's a whole different I suppose kind of different UI but certainly there's a lot of work gone into the basic user interface. And then we've got data flows. We've got the explain feature. But just give us a bit of an overview of what's happened in terms of the UI

and the general, I suppose, things like search and so on in there. How has it changed in this release? What's new? Yeah, you're right. It is a massive release in terms of a number of new features coming in, such as machine learning, which I think we'll talk about in a moment, and data preparations through data flow. But from an end user perspective, what we've tried to do is to make it really easy to learn how to use the product. So we've got a new homepage.

So when you fire up the OAC or DVD, you're immediately presented with, hey, this is how you start using Oracle Analytics Cloud. And you've got links to tutorials. You can watch a video overview as to how the product works. And then the new homepage actually will adapt to how you use the product. So, you know, you've got your kind of initial view, which is going to provide things like, you know, you can use the sample data set, this is how you connect to your own data sources, this is how you kind of import files. But then, you know, over time, it will adapt to how you use the product.

So, you know, maybe projects that you use, you know, a lot will start to, you know, appear at the top. So it's all about intelligently presenting the end user with what they need to know. And, you know, we've taken some of the experiences that we have from our mobile products that's called Day by Day,

which, you know, has what we call a smart feed. And that smart feed will learn about how you as a user of Day by Day use the analytics, you know, whether you're using the analytics, you know, when you've spoken or are speaking to a particular contact that's called you or whether you want to see a certain reports or analysis when you're in a certain location so you know it kind of learns about the end user really and you know this term adaptive is uh quite you know it's all over the place these days in terms of the application of um machine intelligence to the end user

experience and how to make it tailored to that particular end user so there's been a lot of improvements around uh those capabilities and you know the initial user experience for Oracle DV, in addition to kind of a refresh, really, of the UI in terms of the color schemes that are used and how the workflow through the tool goes from preparation through to visualization through to presentation. Okay, so there's been a bunch of new features in DV that came along since I, I suppose, in a way, finished consulting around Oracle. And I think the first one was data flows.

So DV has got a basic data wrangling, data preparation feature in there. Tell us about that and how it works and why you put it in there. Yeah, so the data preparation is very powerful from an end user perspective. So as you say, it's evolved quite a lot since you probably used the tool, which I think was probably at OpenWorld last year, right? I remember talking to you about the data storytelling aspects. And so in terms of the target user, so I remember being at a conference with a colleague. This colleague worked on the business side,

just happened to be sitting down at breakfast. And, you know, he started telling me how great the data preparation capability was. So, you know, he had previously had to jump through hoops to be able to bring data in that he had maybe on a kind of budgeting spreadsheet and incorporate that with you know data from corporate warehouse he could then do it do it himself he could start to build in the logic to join those two data sets together and build up a stepwise sequence for putting it ultimately the data set that he required to do his job.

So that is really what data flows are about, is being able to give the end user the power to do what they need to get the data they need to do their job. And in this latest version of the data flow data preparation capability we have you know added a whole bunch of capabilities for um you know you we've always been able to do basic operations like join data sets filter data sets um as you know as you're aware you typically data is not going to be um perfectly joinable on the first cut so So the ability to edit and transform columns so that they are able to be joined, being able to filter the data.

In this new version, we have the ability to create an S-based cube. So if you're using our OLAP engine, you would be able to connect to an S-based server and create a cube from your data flow. And what's kind of also quite exciting from an analytic perspective is the ability to add groups or bins in an interactive way through the data flows and do that within context of this data preparation step.

And then I suspect we may be talking a bit more about the machine learning capabilities. The data flow capability also gives you the ability to interact now with any machine learning models that you have. So you can run through a training phase for that model with data flows and then output that model as a trained model which you can then apply back to a data set that you'd actually like to apply that model to so that you can do any predictions or you know forecasting on that unknown data set if you like so extremely powerful capability in its own right really in dv4 so how does that work? There's an Oracle database behind it. So what's the engine that you use to do this transformation

and data wrangling and so on? So, yeah, so in DV4, we've essentially got two options. So when you install, and we're talking about a desktop product that's kind of available today. So when you install that on your laptop, there's two other components that you can install. The first one is a distribution of R. And that's something that we've had,

you know, pretty much ever since Oracle DV has been available. And then what we're introducing with this version is also a distribution of Python. It's called DVML, data visualization machine learning. So that engine will then work in conjunction with the DV desktop product to provide those machine learning models and capabilities. Okay, okay. So where do you, I mean, I guess a question that someone like Kent Graziano or someone, or even Stuart might ask is an ETL tool and data wrangling. You know, if a customer says to you,

you know, how far should I go with this? Do I no longer need an ETL tool? What would your advice be around that? And I suppose that applies to data wrangling in general, really. Well, it's a typical consulting answer, isn't it? It kind of, it depends. Yeah. So would you do data, would you do data cleansing in that data cleansing in that for example would you how

far would you go with that really yeah it's it is a really good question and i think the i guess the time scale of a data set is possibly becoming smaller for its validity. So whereas we've all worked on ETL processes that maybe run once a night, the frequency of those updates has become more and more until we're actually talking about streaming datasets. The data flow capabilities are something that you can build up and you know that you can use them to build up quite a sophisticated transformation and it could just be used for a given project, it could be used regularly. So, you know, literally, it really depends on what you're looking to do. And I think the key thing is, is that you've got that

toolbox to be able to do that now. And that's probably where... I guess it's a personal thing. Yeah. I guess it's also when it's a personal bit of transformation. I think that certainly my, you know, one of the things that struck me working in the more kind of startup-y kind of world now is you never see an ETL tool. You may well see ETL processes that are written using Python or whatever or Apache Airflow or whatever, but you never see an ETL tool. And people don't go to transformation and the data sources does not justify some of that kind of work going on. But I think certainly these data wrangling tools, they solve that problem that problem of you know you've got some analysis to do for yourself yeah inevitably the data needs to be

kind of manipulated and and whatever it that that's that's the scope i think really isn't it and you could almost put as many features in as you want to because actually what it's all about is it's it's about doing it for one person really to solve the needs of one report as opposed to being an enterprise thing. Yeah, absolutely. And I think the workers of the current time period are potentially more technically capable or happier doing these types of data transformations to get something that they really need to to get the sort of answers that they're looking for yeah so so okay the other kind of i suppose headline feature of this release is the machine learning support and you've touched a bit on it so far but the thing that's probably the reason that i i kind of contacted you apart from the fact we had

another guest on there to come one moment, was the fact it kind of, you're able to use this explain feature and you can get it to kind of to say, you know, to explain the data within a data set, but particularly sort of segments and segments are things that I work with a lot at the moment in the world I'm working in, where you've got segments of potential kind of customers

and so on, cohorts and so on there. And it's quite a bit, but DV seems to have this ability to explain and look at segments within data. Tell us about that and what it is and, and again, why it was introduced and we'll talk about how it was done. Yeah.

I mean, it's a really exciting feature, you know, as somebody who obviously works in the space, we're all getting up to speed on machine learning and, you know as somebody who obviously works in in the space we're all getting up to speed on machine learning and you know we're all thinking okay we need to start using r we start need to start learning python and you know i'm sure you've installed python onto your machine onto your laptop and you know while it's easy to download it typically when you get it up and running you may need to download some other packages to support it so one one one of the

things that we're trying to do is actually make it very easy for our users to actually start using the power of machine learning without really worrying about how they install python on their on their machine and starting to you know know, have to worry about maybe writing Python scripts themselves. So what explain is really giving you is this kind of almost diagnostic capability to actually take a data set and run, you know, right mouse click, explain this particular attribute? You know, for example, you know, will my customer, you know, churn to a competitor based on the information that I've got for that particular competitor? You know, give me some information, you know, okay, so start with what are the basic facts that that data is

showing me? And then what are the key drivers around those basic facts and then as you talk about you know segments but then anomalies as well i know are there any outliers within that data set so you know what that's then giving the end user is uh first of all an understanding at a basic level of of what the the relationships are within their data set, really quickly, without them having to actually create any visualizations themselves. The tool is doing it for you. And then you can use those as a starting point for further analysis. So any of the charts and descriptions that Xplained will automatically generate for you, you can then bring into a project and then use that

as a starting point and where this is going really is the ability for the analytic tool to actually do all the legwork that's needed when doing an analysis of the basic data set and then the end user is presented with um you know, here's a summary of the data. These are the key drivers and the things that we found that can, you know, are correlated between the data sets and presenting that within a sort of a natural language way as well. And then allowing the end user very quickly to actually determine what action to take on that data. So one of our kind of key drivers really is to be able to allow users to take data-driven actions through this sort of notion of smart analytics for all the users community. So for analysts who are going to be very familiar

with going in and doing these kind of deep analysis on the data through to the business users, their predominant UI could be a mobile device. They could be traveling a lot. So they could use day by day to start getting these smart analytics right through to developers who can start to build

their own visualizations tailored to their own needs which can then be tightly integrated into the product and it will become essentially a first class object like the other data visualization options that we have but they can build it themselves okay okay so so so the example and it was i think it was like using a linear regression or something to to say this the correlation between this this measure this attribute another one is is is strong or not and i think that's that i suppose that's useful for kind of feature selection and and so on there but i mean again you know who do you see do you see this being something that

the average user will will kind of use or is this going to be a data scientist tool i mean what what's your what's your i suppose what's your intention or objective about how universal this will be you know i would absolutely hope that um you know this would be available for all users to use and i can certainly see it being appealing to anybody who's got a question of their data. You know, being able to start getting some insight into, well, you know, what's the data telling you? And, you know, out of that data, which are the important parameters? It has got to be valuable to anybody really. Yeah, I mean, I remember seeing I remember when I went with the eye couple years ago and it was a pretty kind of hardcore sort of feature to use and to make use of and it struck me how much how sort of I suppose consumerized it's been in this this release of the scene now. Yeah, a lot more you point you just right click on

the data set you say explain it comes up with this kind. It's very much a consumer level. It explains it in a way that's easy to understand, but you can see there's some insight that's been gained there. So it struck me as that was, to me, was the, as you might say, tentpole feature before. So, you know, well done for that. That was really good. And also in addition to that, you know,

for those users out there who are the data scientists and may have a library of, you know, Python those users out there who are the data scientists and may have a library of, you know, Python scripts or R scripts that they already know and they trust or they've tweaked to their own uses, what this release gives them is the ability to actually import those into the tool. And, you know, it provides a way for them to, you know, integrate those scripts with the visualization capabilities that we offer within the data visualization product.

So there's like a whole raft of visual options from basic bar charts through to things like chord diagrams, Sankey diagrams, all those kind of innovative type ways of visualizing data are now available to very easily be integrated with any existing ML type scripts that you may have. So I've been playing around with this kind of idea of using machine learning and using, I suppose, automated insights with a BI tool with stuff that I'm doing with Looker. And one of the things I keep hearing from customers with this is things need to be actionable.

And I remember there was something that you guys brought in with OBI, which was action links and the action framework. Is that a stop? And I suppose, how do you solve this actionability thing with BI? Yeah. Is that something that's still a question that you guys are trying to sort out or think about? Oh, absolutely.

No, I mean, as I said, our strategy really is to enable data-driven actions through these sort of smart analytics. And coincidentally, in this version of DV, we have started to incorporate now some action links within the product so what you have now is the ability to create a navigation to a url action as well as you know navigate between canvases within a product so that has already started to be incorporated into the product

um so mike so again the reason i wanted to get you on the show this week was we've got the UKOG conference coming up soon. And that's something that I think we said at the start, that's how I probably got to know you first of all. And you and I have both been, you know, speakers, I think, every year at that conference for a long time. Yeah, since the time I've known you really. Yeah. So just for anybody who doesn't know about the UK OG Conference, tell us what you typically speak about there and I guess why Oracle feel it's an event that's worth supporting. Yeah, absolutely.

So we would not be product managers if we didn't have customers. Our customers... Not for anyone. Exactly. Our customers are our users. So, one of the key aspects, I think, to being a product manager is really getting to know your users. And what better way of doing that than going to a conference or a user group meeting or a special interest group meeting. So, you know, the UK Oracle user group have their annual conference in sort of end of November, December timeframe. You know, it's almost becoming like the start of Christmas because it's typically held in Birmingham now. and they have the Christmas market there that you walk through from the station over to the conference center in town.

So yeah, it's kind of synonymous with Christmas. But on a more serious note, it's a great way, A, for us to meet end users, talk to them in a kind of relaxed sort of atmosphere, really, about how they're using the product. It's a great way for us to present to our end users, you know, what are the new features in the product? And, you know, in some cases have, you know, how to type sessions, you know, the usual

kind of cliche best practices, et cetera. But, you know, people want to know how to use the product really. And, you know, what they want to learn from other users as well so it's not just listening to oracle it's also listening to their you know their peers in terms of other customers what have they done in terms of their system how can they learn from those and sharing ideas around that so it's just an excellent sort of opportunity really to to to network with the people who buy your product and you know learn from them really as to what they think is good and what they think could be improved so what are you presenting on this year what's your what's your presentation

topics uh so i'll be talking about oac um and also the uh the latest capabilities that we've just been talking about in Oracle DB. In addition, we've got John Haggerty, who's one of our VPs of product management, coming over to deliver the BI community keynote. And I believe you're doing a session as well, Mark. I think I am, yeah. I'm doing a session on, interesting, I'm doing a session on, I suppose, analytics outside of the traditional oracle world so i've been you know spending a bit of time looking at i suppose the startup world using tools other than oracle tools um and and what i suppose what um what what these disruptive

companies are doing with analytics and and and i suppose kind of the things we're doing with data now and so on so but tying it back actually as well. So it's a bit of a tour around some of the things going on outside of the world we normally work in, but how it's going to come back and relate to things we're doing now really. So yeah, I'm doing, I think it's on Tuesday, I think. Yeah, I think a lot of the analytics sessions that we're doing, and as you say, you're doing your session on the Tuesday as well,

Tuesday the 5th of December. Excellent. So you've been going for this for many years. What's your favourite memory of speaking in the past? Well, you know, apart from co-presenting with you, Mark, I think, you know, it's always good to, you know, present at any conference, really, and sort of meet your users.

And as I say, it's always a good kind of atmosphere, and especially with UK OUG, because it's around Christmas, and it's almost like the start of Christmas. Yeah, exactly. And actually, we were talking before we did the session, we were talking about actually when you presented to the first ever, I think it was the first ever outing of OBI 11G. us about that and the the kind of the fact that you had the had the audience in your palm of my hand yeah so that was the uh yeah as you say i think it was probably the first

bi forum right because it was in the was it the grand hotel in brighton we had this room no we had this room at the top and um i think i was on a stupid o'clock in the morning quite early. Thanks for that. Yes. But yeah, so OBI 11G was long awaited and I had an early build of that, which had sort of like the charts that kind of popped up on the screen and so a lot of the new capabilities and i sort of very well very vividly remember starting to do the demo and everyone just kind of like almost jumping on me and saying okay cancel the rest of the sessions just show us this all day so yeah that's certainly something that sticks in my mind yeah i think i think being a pm if the if the audience would be a good sign isn't

it really so uh absolutely yeah that was good. That's brilliant. So just to kind of round off then, how would people get hold of DV Desktop? How do people find out about OAC? What's the resources that are available and trials and that sort of thing? Yeah, great question. So if you want to get a trial on the cloud, to cloud.oracle.com and look for the platform as a service PAS under there you'll find the analytics you can sign up for a trial on the cloud or if you go to Oracle technology network if you google OTN Oracle DVD, Oracle Data Visualization Desktop, you'll then go to the download site where you can just download DVD for Mac or Windows and install it on your laptop, get up and running.

We've also got a great YouTube site. So if you search for Oracle Analytics, there's a load of kind of demo videos. They're, you know, quite short as well. So they're in the kind of time period of a few minutes where they will give you an overview of, you know, how to use a feature. You know, what the benefits of that feature are, especially with some of the kind of machine learning new features that we have. I think you were saying that you watched the overview of the new features of DBA. Yeah, I mean, it's a really great summary. In addition, we also have a number of plugins

for the data visualization product. So you can extend the product with new visualization types. We have a number of examples of, you know, how you can create custom visuals. So, for example, you know, we've got built in kind of like geo capabilities and mapping, but you can also create a custom map layer. So, for example, we have an example of a stadium and, you know, an example of how you can, you know, analyze attendees at a gig or a you know a football game or something um so yeah there's a lot of capabilities on there and i can you know provide the links that you can put in the show notes excellent well look mike it's been brilliant having you on the show um and i'm looking forward to seeing you in in in birmingham in a few weeks

time um and say i mean it's been it's been uh very interesting to see or very, you know, gratifying to see, I suppose, how much the product has come on in the last couple of years. It was good anyway when it first came out. And seeing the stuff in there now, I was blown away, actually, by how much has changed and how much new stuff there is there and how much of a, I suppose, how much of a kind of complete tool it is now, really, for having you on the show. Thank you very much and take care. Cheers. Thanks, Mark. having you on the show um thank you very much and uh and take care cheers thanks mark