Drill to Detail - Drill to Detail Ep.24 'Tom Sawyer Software, Graph + Spatial Analytics .... and why a Tweet goes Viral' With Special Guest Kevin Madden

Episode Date: April 4, 2017

Mark Rittman is joined by Kevin Madden and Josh Feingold to talk about graph + spatial analytics, Tom Sawyer Software ... and why a tweet goes viral

So welcome to another episode of Dril to Detail and I'm your host Mark Ripman and this week I'm joined by not one but two guests for the first time because they worked as a team to put together the demo that formed the highlight of my recent BeWire conference talk over in San Francisco, where I analyzed the tweets that happened around the kettle incident that happened with me a few months ago, where we'll talk about it obviously during the show, but where some tweets about things that I did with a Wi-Fi kettle and the 11 hours I took to boil a cup of tea with it went around the world. And we basically, as part of this presentation, we looked at how that kind of tweet went viral. And we used the software that we're going to talk about and the company that helped me on this presentation to actually kind of look at it graphically and see how it worked. So I'd like to welcome to the show then Kevin Madden and Joshua Feingold from TomSci Software. Welcome to the show. And it's great to have you here.

Thank you for having us Mark, we're very much appreciative. So Kevin, oh it's two of you there actually, so Kevin do you want to start off then by just introducing yourself really and a bit about kind of Tom Sawyer Software because you're quite famous in the world that you're in but a lot of people might not have heard of you. Yes, I'm Kevin Madden. I'm the Chief Software Engineer at Tom Sawyer Software. I've worked with Tom Sawyer Software since the early 90s where we started in Berkeley, California. The CEO is my older brother, Brendan Madden,

and we started off working on advanced visualization using mainly C++ technologies for Microsoft Windows. And we were working on more relational management systems for creating schemas and network management platforms in the early days of the internet revolution. Tom Sawyer was the first graph company probably in the world that we know of. So we've been a firm believer in graph theory since the beginning. And one of the important things about the way Tom Sawyer was always customer focused. We always were listening to the customer as far as what feature requirements were. The power of graph analytics and what we're seeing now in graph databases

such as Oracle Spatial and Graph, or they call PGX, Parallel Graph Analytics, we were doing against in memory, basically, since the mid-90s. They were mainly used for network management as far as like internet routing and network management visualization to keep big operations running, such as Sun Microsystems, SunNet Manager. We had an early product that would support Apple Talk networks and stuff like this. We started doing ER diagramming with customers such as LogicWorks, which created ER, Erwin, which created a lot of the relational database schemas that we use today in relational. And then we started doing a lot of database administration applications with like Embarcadero technologies.

Then you saw a big wave into from the object management group for UML modeling. Basically, it was the three amigos and you had companies like Rational that were doing trying to use graph technology for UML to describe software components. And then, you know, you had a big growth in the software industry between the 80s and the 90s and through the Windows programming environments where the complexity level went very high. So Tom Sawyer's visualization was used to create

like these ER diagrams for creating relational databases, network management platforms such as HP OpenView, these kind of platforms for some of our early adopters of Tom Sawyer visualization technology. So that's how you know Tom Sawyer got started. We basically started from the internet routing to we rolled into ER diagramming, then we went into UML modeling, such as there was a consortium between Tom Sawyer, at the time Sun Microsystems, before Oracle did the acquisition of Sun, and another company called Embarcadero Technologies, which is famous for doing database administration tools, we created a UML modeler for NetBeans. And that was a big uptake for the Java platform. moving from C++ into Java, because it was much easier to write applications that could run on the web as far as reducing the complexity required to build applications.

And Java was an interpreted language, and it allowed it to be run on multiple platforms, because we used to be in a big complexity with compiling C++ code into Java. So there was a big migration to Java, and that's when Y2K hit. So there's a big uptake into Y2K. You had a lot of banks who had older software to migrate, and they needed to reverse engineer software. So that's where this big push into the UML modeling came in

and understanding complex relationships within banking applications. So that was a good time for us as well. And so, yeah, that's how our platform pretty much got started and built, Mark. Wow. I mean, you've been around for quite a while then, quite a while, obviously.

Yes, Thumbstar has been around for quite a while then quite a while obviously yes tom sawyer's been around for quite a while we work with a lot of we were what's known as an oem manufacturer a lot of people don't know that familiar with tom sawyer because we were an oem because we would sell our components and they'd be rolled up into larger applications and resold by you know larger organizations such as like computer associates uh ibm or you know hewlett packard um cisco for instance and we were we were considered the secret sauce and yeah um and so they would you know put us in the about box quietly or stuff like that so so yeah it was it was considered a competitive advantage basically to keep it quiet for them and so we would work with them so probably anyone listening here is kind of wondering what on

earth this has got to do with anything to do with kind of tweets and kettles and and kind of user group conferences and so on it's a it'd be an interesting story i suppose really how um how you came to be working in the area that would make that kind of relevant really and how this kind of story came along so maybe i'll kind of tell a little bit of the background really to how i guess you and i ended up talking and collaborating on on this kind of demo um so for any for one or two people in the world in the audience who who didn't hear what happened last year. I went through kind of a project at home and back over in the UK where I basically kind of digitized and made smart all of the kind of devices in my house and fed all the data into a Hadoop cluster I have running

kind of in the garage here to give myself a platform to do some kind of IoT analytics and IoT work on. And it was kind of interesting. But at one point, one of the things I was trying to do some kind of IoT analytics and IoT work on and it was kind of interesting and at one point one of the things I was trying to do was to kind of voice enable everything so what I've got here at home is one of those kind of early Wi-Fi kettles where you have a kind of smartphone app on your phone you can boil the kettle from your phone which is slightly impractical because you have to kind of like fill the kettle with water first of all so it means getting up going downstairs put the water in the kettle and then kind of

go upstairs back to bed boil the kettle boil the kettle from your from your phone get downstairs again make the tea and so on so the whole thing is slightly ludicrous but but on the day that this kind of happened it just took a bit longer to fix there was a problem with the network and so on and and coupled with trying to get the thing to connect to my amazon echo the whole thing i was tweeting at the time the whole thing took about 11 hours to kind of go and what i didn't realize at the time was that the the kind of the mess the tweets that i put out at the start were starting to get retweeted and and over time this thing was starting to go get a little bit kind of viral

and you start to sort of see i think at one one point it was a Twitter moment and people were kind of commenting on it. And every so often you could hear sort of like, you know, the kind of the retweets going off and so on. And so this whole thing went on for kind of the best part of kind of two days. And it became a bit of a kind of Internet kind of internet meme at the time. But what I did at the time, though, actually, was I actually captured all the tweets and all the activity and all the social kind of media activity around this. And I thought at the time it would be interesting at a later date to go back and analyse this data to try and understand, I suppose, really what made the tweet go viral? Were there particular people or particular kind of nodes in that social network that led to the whole thing kind of taking off? And how did it kind of like look over time and over geography as well? So I put this forward for a presentation for the Business Intelligence Warehousing and Advanced Analytics conference in the u.s that runs in runs in sort

in january and i started working with oracle big data spatial and graph to do this but then what we kind of found was that the very basic kind of data visualization analysis you can do that tool got me so far but then oracle recommended that i speak to kevin because kevin's company or kevin the company that kevin and Josh are working, as you were saying earlier on, you've done a lot of work around network analysis, but also this whole area of social network analysis is an area you guys have been doing as well. Do you want to just give us a bit of an overview of what you've been doing in that area and why Oracle thought you better help me?

Sure, Mark. Social network analysis comes from being able to basically understand the core of a network. Our layout, in order to position elements efficiently on the screen, you have to understand its structure. And in order to understand the structure, you have to run graph analysis to understand what are the core centralities or to understand where you should start the positioning. So Tumsler layout basically use these kind of different analytical techniques to understand how to position the elements social network analysis is done by like say LinkedIn and and Facebook you'll see that quite a bit and I think more and more people are becoming aware of the power of graph analytics they're using them every day but they don't

realize it you might see like on Facebook like oh or LinkedIn you might know this person or you maybe you know this guy because it studies who's connected to who. And then it sees that, oh, you're both in the same industry. Oh, you both attended the same conference. Or you might know his coworker. And they call these games like Six Degrees of Kevin Bacon and reachability. And they were jokes in the past, like,, like, oh, you know, how connected are you to Kevin Bacon? But we do these kind of social analytics like centralities.

You know, I have two distinct groups. I want to understand who's the middleman between these groups so I can understand how to separate them. We use these kind of techniques to hide complexity of the network so that we can aggregate them up and create high-level views and be able to partition the network so that we can visualize subcomponents of the network as well. So it's basically being able to divide and conquer a very complex linked structure and create

actionable intelligence from them um whether they you know it could be a recommendation and engine for instance like uh your friends bought these books and or somebody else in your field bought these books you might be interested in these books as well you saw companies like amazon come up and and use these very similar techniques and they've basically driven some of the uh large part of the e-commerce revolution that's been going on for the last decade or so um when we see all these kind of applications that are graph based like facebook for instance has this their thing is called graph search right these are all built on top of big data platforms and they basically create these

relationship diagrams under the covers and there might be billions of connections and they're on the back end they're running these graph analytic application algorithms to extract out details that you might not know about yourself. So that's the power. And getting back to your question was that we've been a partner with Oracle for over 10 years. We've worked with Oracle over many years. We were actually an OEM into Oracle as well. So Oracle has wrapped Tom Sawyer products into their own products. Their enterprise warehouse builder, for instance, is an example.

And their tools for doing ETL processing. So we've been doing applications with Oracle for many years. So we had a long track record with them. And we also supported, you know, over the years, we supported their RDF graph, which is used by Symantec Technologies for understanding language structures. You see a lot of Symantec work done in Europe for understanding language processing and natural language processing and these kind of things. Text analytics, you'll see a lot of text analytical work using RDF. So we did from relational to their search platform in DECA. You see there's social relationship management systems there uh and the whole goal

is to to basically use these kind of analytical techniques and allow them to create uh bi dashboards so that they can act upon the information contained within these big data stores okay okay i mean so so certainly something that that interested me in working with you was to not only just look at, I suppose, the network at a point in time. So I had the kind of set of tweets and the retweets and the relationships between people, which meant that we could look at, you know, at a point in time in the end, we could see who had commented on this story. And basically, you know, I suppose the full extent of where it had gone to around the world and so on. But I was particularly interested in, I'd seen, I was recommended this tool called Tom Sawyer Perspectives, which was one that was pointed out to me as one that would be particularly interesting for this. Because I was interested to see if we could somehow show over time how this kind of, how the kind of tweet went viral, but also showing it over kind of like looking at it spatially as well.

So but you guys did some you guys did some fantastic work with the demo. Talk us through, just talk us through how that kind of worked or some more background, really. Sure. Tom Sawyer has an advanced web application deployment framework for working with relationships, temporal data, and well, any kind of data really, but we end up extracting data or graphs out of the data. And then we extract out the temporal components as well. So as we're going along, we're extracting out like a time range. So as we took your extracted data and we put it into the graph database we were able to then extract the graph out look at it over time and build a web application very rapidly I think you we had less

than like less than a week it was supposed to be wrong yeah yeah it was less than a week where they asked you to be the speaker at the conference and to get an application running in very short order. So we worked very closely over the years with Oracle. And so we were actually able to say, all right, let's give it a try. I know we have an IDE. We have an IDE designer that allows us to rapidly build applications and build deployable web applications. So we worked with our services department and we took the data set. We analyzed the best way to visualize the application.

We support spatial views as well as temporal views, as well as your standard kind of bi charting pie charts line charts trees and tables and such and we were able to extract out who was the most influential tweeters who who had the most influence we were able to then create a graph model out of the tweet activity. And we have an animation system which allows us to play how the tweets unfolded over time. And you can see them on a global map. It's sort of like Google Maps, but it's our own mapping platform, which allows us to basically plot the tweet activity across the globe using the spatial coordinates from the Twitter API.

Some of the tweets had spatial coordinates like lat long information, which allowed us to plot them over the globe. And you could watch it bounce between the US and Europe. And then it went to Australia and it went, you can watch how over time it propagated itself. And then you could go to our graph views and run graph analytics to calculate who were the central actors. And Mark was obviously a central actor.

And then you also started to see Aaron's Cafe, which is a central actor. And then you also started to see Aaron's Cafe, which is a famous cafe. And then some of the newspapers, the Guardian newspaper. So we were able to play that over time and watch it unfold and then run our in-memory graph analytics platform to extract out who was very influential in making that tweet go viral. And so from concept, from initialization to deployment,

to Mark standing on the stage was less than a week. And there was some back and forth, but, you know, we pulled it off. And I'd say, Mark, you did a wonderful job. You're a very charming man. You got a standing applause, and I think it was fantastically good fun you know it's fun to do this kind of stuff and it it's it shows the power of the of uh of our platform and it shows the power of uh social analytics um and I think with a little more time, we could have done even a better job. But there was a lot of tweets and retweets. And I think there was, I can't forget the size of the data set, but it wasn't small by any means. It was interesting. I think it was interesting

you and I building it because it was a good data set. But also as you build it, you realize if you collected more data, it could have been even more interesting. And I think that one of the challenges a little bit with the data set was that it was largely kind of one-way conversations. So there'd be people kind of referring to me or referring to The Guardian for quite a while. And I guess they're quite hard to kind of model

in an interesting way. I mean, what would have been interesting also would have been to understand the conversations that go on before um this happened to actually maybe maybe find maybe if that could have been a good predictor of the paths that the kind of the the the uh the conversation could have gone but but certainly it was it was interesting to see there were certain kind of nodes in that in in that network i remember at the time it was there was a sort of i think there was charles stross who's a kind of a science fiction uh writer over here and there's various kind of people when they mentioned it you could see kind of suddenly there was a kind of you

know suddenly the acceleration kind of went off the scale with the kind of the network and and the kind of the you know there was that node was obviously very influential in that network correct um i mean with that much information it's very important to understand who your was obviously very influential in that network. Correct. I mean, with that much information, it's very important to understand who your influencers are and who you want to build into your networks in order to have PR impact. If you're trying to have an impact with a tweet, I mean, people tweet because they want attention.

Thank you. No, I mean, it's a hilarious topic. I think you were showing off the IoT. I mean, that's a big field for us coming up. Here is IoT. We're still in the infancy of IoT. And you see a lot of big data applications where the amount of information coming in

from these distributed network of devices, you have, you know, you have security challenges, you have massive amounts of streams and RF and you have a lot of complexity. But at the basics, it's meant to be simple. Like I need, I need milk, you need milk in your fridge or your refrigerator is telling you you need milk. But it could be as, you know, important as, you know, you know, are, you know, are my wings going to fall off my airplane? You know, these kind of things. Is there, you know, is there fuel in my car? Is there, you know, kind of oil fields and nuclear power plants. And as we start to get into more infrastructure-based IoT applications,

it's going to grow in importance, and you're going to see more importance around network security and distributed network management. And this is where Tom Sawyer's perspective really shines because we've been a network management platform for a long time. We allow you to work with both relational databases and NoSQL stores and graph databases and federate them into a single view. But you don't have to be a web expert. You don't have to know how to create all that complexity. We try to boil it all down and focus on what's important is

what do I want to draw on the screen with this information? What are the connections between these two elements that define something I want to understand? And that's where perspectives really shines because we have a rules engine that allows you to create, we call it perspectives because you can create multiple views or different views on that same data set. So one case we wanted to look at it spatially. And in another case, we wanted to look at your tweets in how it was interconnected and how those influencers. So basically what we did is the more influential you were, we basically used the number of followers you had to decide the size of how big we made the nodes. And then we basically used them to create weights, and that they had more weights

in what they say. If somebody who's very influential endorses you, that's more important than somebody who's not important. So it allows you to separate the wheat from the chaff, and without being a web expert, build advanced analytic applications that you can use in your everyday life, whether it be for a small work group or for company-wide deployments. So you guys have quite, I suppose, a different approach

to data visualization and analytics and so on to what I think is more traditional kind of bi and bi vendor tools i mean it's interesting to think about oracle have that kind of that that pgx engine the in-memory engine there you've talked about graph analytics and so on but you guys you know especially with data visualization your your kind of take on on kind of you know visualization graph analytics and so on it's quite different isn't it really it's very powerful and it's a it's a kind of very um it's definitely a very kind of like different take on things and potentially a more powerful way of doing things than just standard graphs and tables isn't it uh yes it is i mean oracle and the databases are very good at dealing with the billions and

the trillions and the terabytes and the petabytes let let the graph analytics and the databases are very good at dealing with the billions and the trillions and the terabytes and the petabytes let let the graph analytics and the database engine do its job over the bigger data sets right they they have the power to crunch those kind of things you still come out with a graph to the client like if you query a graph database you get a graph out most of the times and so you're but what you're doing is you're using graph analytics to generate a smaller graph, which is then presentable to the user. But you still have multiple dimensions you want to include into that. And you also want to tie other sources, like for instance, the Twitter API. You know, like in one case, you want to have some relational store that you keep other details, and then you might have a graph database keeping the part that understands the links, how you've linked things together. But then you also want to augment it, let's say a web search.

So all of a sudden, you have several aspects that you want to pull together, and you have to do it on the UI level. But then you still want to do graph analytics on that so that's where tom sawyer really shines and and once we reduce the complexity to getting an application to the web it and you tied like features like animation um and push events and all these kind of things which allows you to understand to drive the ui from data so that that's some of the secrets of tom slur too is that everything is data driven and so the ui is very responsive it's in real time and it allows you to aggregate multiple sources together and still run graph analytics yeah that's yeah i mean you're being slightly

kind of you're being slightly kind of you're being slightly kind of uh what's the word uh underplaying it there actually because the reason the reason that you came in at the last minute to this and to help was i think i spent about the previous six weeks just trying to load the data into oracle big data spatial and graph yeah and i think that that to your point there about about tom sawyer perspectives being easy to use, graphical, all these kind of things there, the progress you guys made in just a week was incredible compared to what I've been involved in. That's not to knock the Oracle software, because the Oracle software is much more basic, and they've quite quickly said to get you guys involved.

But the speed and just how quickly you got something very kind of visual and very useful together was fantastic really and really you guys saved the day really when it came to the actual presentation uh thanks mark well yeah we've worked very hard over the years yeah we we we do have something special i believe that tom's perspectives and you know our team of engineers has been working diligently for years um to pull a segue we've basically uh gotten kudos for that like of staying narrow focused on what we're very good at stay laser focused on graph visualization graph diagramming uh web application deployment, and stay laser focused on what the customer requires.

And don't try to wander off into other kind of, like, you know, we could have grabbed one of the open source graph databases and tried to bring it up and become a graph database company as well. And we've gotten high marks and high from other people that said we're so glad that you've stayed laser focused on what you're great at yeah and i think that's it's been an important differentiator for tom sawyer yeah yeah definitely to tell us about i mean so the other part that you guys helped me with was with the spatial side and so when i put this up when i put this conference abstract in it was very much going to be about not just analyzing the network as in a network kind of graph thing

but to look at it spatially as well so tell us a bit about how you did the spatial side and and i suppose in a way a wider question or wider thing how can spatial analysis be interesting in this sort of area as well uh spatial um uh the importance network, like we've been dealing with network topologies, and networks have locations. Usually they have positions. Tom Sawyer has created a patent-pending technology which allows us to render network topologies on top of you know maps on top of maps and it allows us to zoom down and see more information that like say you want to see the network inside of a building so you start at your google earth view and you zoom down you want to start to see more and more that network information this allows us to like have a global view view of things and then get down to, okay, hey, my remote office in Japan is having a problem. Let's zoom down and see what's going on there. So it gives the IT departments the ability to understand departmental structures across the world and how they're interrelated. Process dependencies, you see like industrial production, industrial, you know, for instance, our customer Airbus and they manufacture fuselages in another.

And then they have very complex processes between all these production facilities. You know, there's like five million parts in an Airbus A380 and trying to keep track of all these kind of things. Really, when you put it on a spatial area, you can look at it from a spatial perspective, or you can also look at it from a logical perspective. Like, I don't care where the network is placed. I want to look at it like how it's what we call a topological structure and how it's really interconnected. So there's two different views to the same kind of information. How it's wired together is sometimes more important than where it is physically located. But sometimes you want to look at it from either view.

And you can do spatial queries, which might give you a road network, for instance, and understanding how I can route my car through a road network safely. People use it all the time. For instance, they use uh car routing software to how to get to work oh there's a traffic jam you know so how do i avoid the traffic jam those are graph routing and graph analytics problems that they that they use every day but they just don't know that they're doing this kind of things yeah i mean if you think about i guess a very topical kind of way that i imagine you guys are getting involved in this if you think about with things like brexit for example um you know you think about if you're a car manufacturer and you're manufacturing, say, your Nissan, for example.

And you're there may well be a customer of yours. I don't know. But, you know, you're Nissan and you've got you've got you've got a factory in, say, sort of like England. And you've now got kind of potentially Brexit coming along we got Brexit coming along and you're going to find there that some that that you've got potential kind of now you know customs duties across kind of you know from the UK to Europe and so on there and you've got it you've got a supply chain and you've got a network of kind of parts you've also got the kind of that you've got geographic element as well there that I would have thought would be absolutely perfect for the kind of things that you do I mean also with things like NAFTA and so on anywhere now when trade barriers are starting to

go up and and but your supply chains are so complicated i imagine this is exactly what you do really yes that's correct i mean some of our customers like the jet propulsion laboratory are doing very complex systems engineering we're doing a lot in the aerospace industry, and they're multinational, and they have complex diagramming in production facilities. They have complex diagramming requirements inside of understanding how systems are integrated together. So basically, we have these very complex systems built on top of other complex systems. And I think this is a great place for my coworker, Josh. He's done an excellent job working very closely in the aerospace industry,

applying solutions and how Tom Sawyer Solutions basically take our Tom Sawyer Perspectives product and how we deploy to our customer base. Basically, that's how Tom Sawyer engages. And most people are like, well, how do I engage? Most people start by looking at a small data set, and we start with a small data set, and then we apply it to the much bigger. But I'll let Josh Feingold, our lead solutions engineer, discuss some more of the work that he's been doing in our solutions department. Good, Josh, nice to meet you.

So, yeah, tell us a bit about what you do then and some of the things you've been doing around that area. Oh, yeah, sure. So my job title is lead solutions engineer, which is kind of a job title that doesn't mean very much. But what it means is that basically, whereas Kevin is kind of the ruler of what goes on inside of Tom Sawyer. As far as engineering goes, once it starts going out the door and becomes customer-facing and is owned by the customers, then it kind of enters my domain.

So I help customers who are interested in building applications, either with training or coming in and having our services department uh actually build up applications for them or or teaching them to use uh the tom's area perspectives tools okay okay so so give us an example then of i think you mentioned aerospace there as well i mean what kind of what kind of projects that do you do then uh that would be that would kind of projects do you do then that would be a good showcase of this, really? Give us an example of one you've done, then, really.

Sure, sure. Now, we have a lot of growth, actually, in the model-based systems engineering community. That may not be terminology. So what's that, then? Yeah, yeah. So model-based systems engineering is basically, if you imagine a company and you've got to build a new product. So let's say you're a paint company and you want to start a new line of paint. Well, that seems like a relatively simple task, right? You've got to come up with some colors and you've got to come up with some base paints. But that's actually an incredibly complicated process where you have to gather requirements. What's the paint going to cost? What's its reflectivity going to be like? What are all of these chemical attributes? What are the compliance? Where are you going to have it built? And so you have to start this process of gathering requirements. And then you transition that into an approach and you come up with physical

systems that will address this and then physical systems that will distribute it. And now you have to be able to verify that all of your requirements are actually being met by your systems as planned and that your systems is built, match your systems as planned. You know, you can see that this is fundamentally a big data application. And so what goes on in there is that for years and years and years, I mean, for many decades, this has been a document-based system where someone would write up a proposal and then they would, you know, print it out and they would walk it to the desk of the person, you know, who needed to sign off on it. And you would end up with this essentially stack of papers that, you know, people had signed off on. And now in 10 years, when it's time to change a decision, right, because you need to revitalize your brand or something, or you've got inefficiencies somewhere in your manufacturing, now all you have is the stack of papers and you have no chance

of getting your data back out, right? So model-based systems engineering is taking this standpoint of, I'm going to model this whole system end-to-end. I'm going to keep it in some kind of database. I'm going to use some kind of language so that when I come back in 10 years, we have the whole thing end-to-end and we can see exactly what the criteria were. We can see who ran the tests and it's all continuous from end to end. And it's not this kind of like broken up, fractured set of documents that were authored by people using human language instead of kind of machine readable logical constructs that can be reconstituted. Yeah. So where would, so is there, is there a kind of a scope for data visualization in that? Is it something where...

How would people kind of visualize this and work with it, really? Right. So historically, there have kind of been two camps. And the first camp is people have sort of rolled their own. And this is actually when Kevin mentions Airbus, Airbus Gaia application. That is an example of someone coming up with needs and just rolling their own. They said, we need this. And so we went in there and they said, well, you know, it's great that we have all this data, but we need to interact with it. We need to visualize it. We need to be able to convey

this highly technical data to people who maybe aren't highly technical, right? It needs to have an intuitive front end. And so we go in and we deal with the exact needs of those types of customers. The other side is that there is a standard for this called SysML, the systems modeling language. And we also work directly with the standard. So I'm actually one of the data visualization co-leads for SysML 2.0, which for people who care about such things will be releasing its RFP at the end of this year, and we'll probably be reviewing and accepting proposals another one and a half to three years after that. Right.

So, you know, it's the systems modeling language. And basically, it is both a definition for an abstract syntax, right? So a data model and a concrete syntax, meaning diagrams. And then we take that data and we build it into automated diagrams and maps. And yeah, so if you imagine in your Twitter example, what if you had this Excel sheet with thousands and thousands of tweets on it, right? And then you had to take your map of the world, and then you wanted to put a pin in the physical location where each of those tweets happened.

And then when you realized that some of those tweeters were more important than others, you wanted to replace your small pins with big pins, right? Imagine you had to do that by hand. Oh, yeah, yeah. That's essentially the state of the industry yeah exactly i mean i mean my example was massively trivial compared to what you're saying there but it was when you i mean even just building out myself to start there's actually a lot of dimensions to it a lot of things to there's a lot of kind of aspects

to it really but as you say doing that kind of on a on a kind of like a larger scale and so on it must be kind of crazy but it's no, no, I mean, it sounds, the reason I wanted to speak to you guys was because I've always been aware of you, you know, Tom Sawyer, through the kind of user groups and through kind of your attendance and participation at the BWA event. I was kind of interested. It seemed like you were kind of a company that not many people had heard of outside of your industry, but were kind of doing massively interesting analytics work and stuff that was very complementary to the kind of the the the um

the kind of you know the graph the graph features really in the oracle database and oracle big data and it's um it's been kind of interesting hearing kind of you know what you do a bit more detail on it and so on um so so just just to kind of if so if anybody people would be interested hearing this i mean how how would you how would people engage with you how would people kind of, so if people would be interested in hearing this, I mean, how would you, how would people engage with you? How would people kind of, is it either the customer or is it a sort of developer? What's the kind of route in to understand more about kind of Tom Sawyer and kind of the graph analysis that you do as the kind of backdrop to all this? Oh, sure.

So the, you can come at us from any level. So we have OEM customers who want to come in and they want to get all the way down in the weeds and they don't want to have us do anything. And so we offer a full rich API. Then you have companies like the Jet Propulsion Laboratory, like Airbus. We have multiple Fortune 500. We have some Fortune 20 consumer goods manufacturers, top global auto manufacturers. And what they do is they're not interested in being their own developers. What they do is they say, hey, we've got this need. Tell us what you can do for us.

They come to us and just like we did for you where you had this massive data set and you wanted to see something that had to be ready to present in a week. Well, when you have more than a week, imagine what we can do, right? I know, exactly, exactly. Yeah, so we do professional development. And then we have in the middle of the two, we have training and integrated development where we'll come in and we'll train you. So no matter where you are on the level of, you know, I want to be down in the weeds at the API to I just want to get something that looks stunningly beautiful and does exactly what I need to do to make my job, you know, 10 times faster and 100

times easier, you know, that anywhere on that spectrum, we can work with you. And so to Kevin, really, I mean, if you're if you're an Oracle developer or big data developer, and you're looking at you, you present at a lot of of conferences and you're engaged in that community quite a bit. How would maybe a developer kind of find out more about Tom Sawyer and what you do? So if you're coming at it from a relational point of view, Tom Sawyer easily connects to your relational data stores and we extract your tables and your foreign and primary keys

and we allow you to basically move the visualizations and they're powered by queries. So you might do select star from and grab some data. Now I want to define relationships between that and then just get it to the web. I mean we've been a long time in the relational space and now you're starting to see this big move from relational towards the graph database and towards other NoSQL stores. Your Cassandra's and your Neo4j's and your Oracle Big Data Spatial and Graph, because they're starting to understand the power of these analytic applications that can be run on these linked data sets. So if you're trying to come at it from a relational point of view or you're trying to migrate from your relational store to a NoSQL store, Tom Sawyer Perspectives is a great way to go

because we allow you to preview the graphs prior to moving from relational. So you can create like a process, a migration process from relational to graph database. And once it's in the graph database, we can go one-to-one in memory with the graph model. So it's funny, over the years, we use graphs to define relational structures, your data modeling language, and your entities and your relationships

between the tables using foreign and primary keys. And then we would create the graph we would create the database and then the database would do it in relational relational ways using constraints and this kind of things but they were also heavily used in basically administering database applications like creating alter tables and all these other kinds of we did some a lot of work in table space mapping oracle defragmentation but so there's no fear in moving to keeping your data in relational and then moving parts of it to the graph database that need to do the analytical

parts for graph database so you don't have to you don't have to really remove all that or or or oracle data or relationship relational data you can move only the parts that you need and you can federate the two together to create a seamless view between these applications and that's the power of uh federated uh data integration model good stuff good stuff well just to kind of wrap things up really i mean you've obviously you built out the demo and i think you've sought a copy of it and you show it at sort of various uh you've shown it some events whereabouts you guys kind of presenting next is there any way people can kind of see what you're doing and maybe sort of see an example of this demo and see how it kind of works um yes uh hang on a sec we're uh you can go to tomstutter.com and you know you can use visualize the panama papers

application like we did a panama papers uh visualization we do some link analysis applications network management apps if you just create a login on tomstutter.com uh we have a very active schedule uh just last week we're presented at the omg conference and then uh two weeks before that we were at the big data paris uh trade show but we do uh we have the bwa conference not the bwa well oracle bwa is all we were there every year but uh we go to the oracle open world shows and we also do uh um we're doing the gEO-IN spatial conference, which is going to be in Austin or in Texas. And then we're also presenting at the NoMagic, which is a system,

NoMagic is a modeling framework for dealing with like both UML and SysML and other kinds of diagramming. I'll let Josh talk a little more about that quickly. Oh, yeah. So we're going to be at the NoMagic World Symposium. And NoMagic is probably our top partner in the world of model-based systems engineering. these tools that's been allowing, allowing customers to, to build their giant complex models for, you know, I think the past 20 years. And now we've come in and taken the manual step of, you know, arranging all of these diagrams by hand and making them pretty by hand that

used to be very time consuming and and so with uh no magic we automate that so that it takes you you know a couple of seconds versus a couple of hours uh to get something that looks like beautiful enough that you can show it to your ceo and he won't tell you that he doesn't ever want to look at you on that again excellent excellent well look it's been great speaking to you it's really glad to have you on here and nice to speak to both of you. Thank you again, Kevin, for the help you gave me with the demo and the conference presentation back in January. And Josh, it's been great to speak to you as well.

Thank you very much for coming on the show and have a nice day. Thanks a lot, Mark. Okay, take care.