 Welcome back to Strata Live here at theCUBE, SiliconANGLE.tv's flagship telecast. We'll go out to the events and talk to the smartest people you can find, interesting people, smart people, and we extract the signal from the noise and share that with you. SiliconANGLE.com has all the coverage. SiliconANGLE.tv has all the videos on demand. This is live and we're going to talk with data scientists here. I'm John Furrier with FoundersSiliconANGLE.com and I'm joined by my co-host. I'm Dave Vellante of Wikibon.org and we're here with Jesper Anderson who's the founder, general manager of Bloom Studios. Also, as John said, a data scientist. Jesper, this is one of those deals where a picture's worth a thousand words. I mean, you just go, for people in the audience, Google Bloom Studios and check out the website. It's gorgeous. It really is. And, you know, I'm used to it. We all are, you know. I hope they don't have the cube effect because the cube effect is when we give a website out it just crashes, so, you know. But, I mean, you know, when we prep for these calls we go to websites, we learn about products and things like that. I went to your website and went, wow, I've never really seen anything like this and I know Mark is going to bring up some examples but before we do that, why don't you tell us a little bit about Bloom Studios, where you got the idea for founding the company and what you guys are trying to do? So Bloom Studio is essentially charged with making consumer applications out of data and data visualization. So we're trying to make toys of your data that later on become tools. And the idea is that sort of visualizations mimic, the interactions with data visualizations mimic the way you interact with video games. And so that we thought we could take those mechanics and layer them on top of each other and make something that was really pleasurable to use so that you could peruse your data in a more engaging way. And so one of our iPad apps makes a galaxy metaphor for your music. And so instead of just looking at the list view of artists, whatever, you navigate down through solar systems and planets and moons to get to the actual tracks you want to play and just kind of go through an experience that's much more reminiscent of like a Mass Effect game or something like that. So is the primary motivation for using your tools just fun, it feels good, it's cathartic? Is it finding other, what's driving? So I think the first motivation is definitely pleasure. It is to make something different that's more engaging. But then visualizations support serendipitous discovery in ways that's more direct search or this views don't because you have to have declared your intention before you go to look for something. And visualizations tend to bring forth relations that didn't exist before that you wouldn't be aware of and you can discover. So by having these beautiful visualizations you're finding that people are discovering new information and new ways to interact with their data? Absolutely, that's why the field was invented in the first place. So we're just doing that more front and center. So we love rich media, we do the live broadcasts, we capture on demand, we love capturing content, we do some photos, but media's sticky, photos in particular, what are you guys finding in terms of the elements of communications? Is the user experience one kind of combinations? What are you seeing that resonates best from a visualization standpoint these days? Because we all know Facebook has a popular app called the Photo that people post in zillions of photos on Facebook. What media is resonating the most out there from your standpoint? Photos are definitely key for us. I think that it's really, there's a lot of metadata. So there's around photo applications. So there's a lot you can do to kind of expose new relationships between the data months and it is highly visual. So it's really interesting. One of the things that we're really interested in looking at next is video and seeing what we can do there. Where it's a little more complicated because there's less metadata for the amount of content available. But at the same time, we think that it's a huge problem that there is this sort of conundrum between the armchair experience and the TV experience that we want to try to bridge and we think that data visualization could make finding new snippets to watch faster and easier at a more pleasurable experience. Pinterest and Palo Alto where I live is a hot company and you got all these new kind of techniques. I guess my question I have for you is more of a philosophical one based on your experience and maybe vision around the future user experience. The web obviously evolves from static web pages too. The user experience has always been a cool area to push the envelope on. Now with Cloud Mobile, socially you have analytics, you've got data science and you have a user experience now that's have different elements to it. You have mobile, you have TV, has been connected to Netflix soon to be bundling with Comcast, all kinds of new things happening. What's your vision around the user experience and their expectations of the users? I think what we're trying to, the way we think about what we're trying to get is maybe going more to the original sort of unix like metaphors for the web and trying to think of more sharp tools and how can we combine them so that you start to think of the tablet as the moving more to HTTP and not HTML and the tablet becomes the way you view or tablet applications from the way you view the web content and then you can chain those things together in more interesting ways and get back to a sort of a tool set instead of a one-month monopolistic experience. Can we get an example? No, because we're inventing those right now. When you say unix toolkit experience, what do you mean by that? Do you mean from a developer or from a user? From a user standpoint that you might use one application to find a set of tweets, for example. And those tweets, you might want to select just the music ones and then take those music tweets and pipe them to a music application to play them in one experience or to a video application conversely or something like that or make sharing a little more direct so don't share within the application but take the data you want to share into another application and then share it from there. So there's a couple of tools that Mark, I think you have these, you can bring them up. So one is a cartogram. It's essentially a way to show public photos. You type in a place and it'll show you so I can type in Santa Clara, California. Great, we should take you to what was interesting in Santa Clara recently. And it'll show me what's happening in Santa Clara. How's it going on in Santa Clara? Santa Clara is very interesting about Santa Clara. Great America is right next door. Maybe I type in Italy or something like that and then it shows me the visualization of all this cool data. Now where's the data come from? So this is all from Instagram. And so we're using their public APIs to get as much information as you can for all the different geo points. And then we are behind the scenes doing some graph calculations to figure out what are the most important photos from there. So not just using their popularity metric but coming up with one of our own that's more about significance. Okay, and without divulging the dark secrets of your algorithm, I mean essentially what does it do for me as a user? You can really think of it very propitiously. It's page rank for photos. And basically it just takes the users who contribute the most in terms of voting and takes their votes and makes them more important. And then Fizz is a way to visualize Facebook and Twitter data. It looks like you got linked in on there as well, right? Yeah. So if we click on the Twitter, I'm gonna do something wrong. So I think he ends that. It's gonna take a little second, but I'll come through. I've got a slower connection than we probably do online. So while Dave's doing that, I want to kind of ask you around, because we think about, first of all, we love what you do. You guys are doing fantastic, but big supporters of it. Thank you. We're challenged here at SiliconANGLE and Wikibon because our goal is to provide as much content as possible in a social format and make it open source. We don't charge for any of our content. But we have a website, but we're not trying to build it as a destination. We also have video. We also have photos. We're constantly challenged around the design problems and challenges around getting away from the locked-in Walt Garden website model to something more promiscuous and frictionless, where the user experience can be better than some of the other content sites out there. What should we be thinking about? Because this is not just our problem. This is everybody, right? I mean, everyone who's in the media business has to eventually get to the folks who are natively, what I call the Apple natives and the mobile natives, like my son is 16. He doesn't read the New York Times, never will. He'll never really watch TV much other than gaming shows or anything he's interested in, maybe. So there's a whole new set of generation of users coming in that need to access content. What should be people thinking about? What should we be thinking about? I mean, I think, again, it's going back to applications that there's a richer experience there on the apps, on the iPad or on your desktop, and that the website itself is becoming a little too brittle and a little not flexible enough to do what you want to do. So once you're within the app context, then you actually re-own the experience, even though it's a micro experience, and you can start to do things there to both reward yourself and your users and create a more controlled but flexible experience, as long as you give your users a way out, which is really the angle that we want. Well, make it free, or make it available. Free the lead, not locked in. Certainly make the content shareable and findable on the web. Use the website as the way to share and what path that is actually, to do a really good job of the path application is this lockdown experience, but all the elements are on a website somewhere that you can share. So I have this other application, FizzUp. I don't know, Keen, if you were able to log in. So basically we've got these, it's basically a representation of Twitter data from 107 people, for me anyway, with 200 updates from about 50 minutes ago to about 26 seconds ago, so near real time. Yeah, near real time. And so I've got these sort of blobs and there's a huge blob in the middle, so it's at furrier, what's that all about? No kidding. Dominating my Twitter stream. And so all right, so what are we looking at here? So this is essentially that every large circle is a user and every small circle within that is a tweet that that user has sent. And so this is a way of kind of breaking free of the timeline view that sort of predominates all of the Twitter clients. Yeah, so you can kind of get back to like, maybe you took a break from God forbid from Twitter for a day or something like that and you can come back and get a cross-section of what happened, what was important and who was talking about what. And you can type the keyword in the search field, maybe you can see what was a strata activity today or something like that. Yeah, so if I do that, if I type in strata, it's going to narrow the search down to those tweets that are relevant to strata. Exactly, just those tweets. And then you can kind of see, but also the users that provided those tweets, you can get a sensor who was loud today about that, who had an opinion around that sort of topic today. So the size of these bubbles didn't really change, but the data inside the bubble's changed. Yeah, absolutely. I realize that, you know, Stevie Chambers was maybe even more active than Techie, Junkie and John Furrier was very, At Furrier was very active, et cetera. Exactly. So what kind of, so I'll get back to my questions. I mean, maybe, I mean, obviously personal agenda with making our site better. Is there a technology that we and others can use? Like, are you guys selling the software? What do you guys, I mean, we need to figure out a better CMS and we're always constantly looking at what can we be programming in. And also we have a big data backend that we use. I mean, we've been really, really successful with Hadoop and HBase. Some of the things we're doing on the data science side for SiliconANGLE, Wikibon. But our web presence, it's, I want to make it better. Is there a better, we use WordPress, right? WordPress is obviously open source and cool, but I want a better C, I need better software. What can we, what is out there? So, you know, for the CMS thing, we rolled our own too. We didn't really like what was out there. But on the front end, you're seeing sort of a novel, like the growth of these novel frameworks to take advantage of what HTML5 can do and what WebGL can do. And those frameworks are starting to come into maturity in a way that actually is accessible to people. And so, Mike Bostock has some really great work first with that front of his now. Who did? Mike Bostock. Yeah. I think he had a presentation earlier this week. But, see, this is great sort of dynamic web page. Library called D3. And you can see it's something else called 3GS that's a. Mike Bostock? Bostock, yeah. Bostock. Is it square right now, I believe. And then there's another library called 3JS that sits on top of WebGL and makes that accessible to mere humans and takes care of all the boilerplate and things like that. And I think those frameworks are going to start to be the ones that become a lingual franca of the kind of next level of really interesting website. And certainly the ones we're leveraging when we build websites now. One of the things we think about is obviously iTunes and Apple, right? Tablet, huge, and Android. What's going on there? I mean, what's the market like over there from a tech perspective, from an integration as people move content into there? Is it just straight HTML5 they got to worry about for Apple or what are some of the things we're finding? We code almost exclusively in C++ for Apple in order to keep it marginally portable, though we're still not completely compatible with Android, unfortunately. As for HTML5, for our perspective, because we do a lot of dynamic, three-dimensional stuff, we're somewhat hamstrung because they haven't opened up WebGL on the WebKit browser that you can use natively, which is only, you can get that for iAds only. And so we're waiting for that before we really leverage HTML5. However, we do really enjoy to be able to create a WebKit and that's how we show Twitter content or deep content whenever we can. Yeah, what about the backhand? Obviously, what kind of big data infrastructure and software using Hadoop, HMase, Couch? We're actually exclusively on Mongo and using streaming analytics, so we try to be scale-free. Which analytics? Streaming analytics, so we calculate all the analytics on the fly and try to make it be scale-free that way. Cool, anything else? We're going to just, it's Node, MongoDB. Node.js? Node.js, yeah. We did the Node Summit. I love that community, man. I think Node is probably one of the hottest things happening right now. It's definitely where all the hipsters are at right now. It's good work coming out of there. I mean, it's not just hipsters, it's dudes pumping out some good code. Oh, absolutely. My O is great. Your background, Jesper, is that data scientists is interesting. I'm just noticing you've solved some problems in financial services, credit card fraud, home valuations and the likes. What led you to this sort of visualization? Well, I started getting more interested in consumer-facing data science and what we could do. It sort of, doing the home valuations was definitely geared towards how we'd present users without the home valuations and things like that. And confronting sort of that, as data scientists culturally, we've been so accustomed to working exclusively on a B2B level and providing actionable business metrics to users and that all of my training wasn't actually very well geared towards providing consumers with equivalent sort of services. So Bloom was sort of a good chance to kind of stretch those muscles and wrap my head around and wrap our heads around what it would do to how you would present data science to users in a more straightforward way. And it turned out to be quite different answers, we think. Upstanding. Okay, so thanks for coming on the queue. We really appreciate it, Jesper. You guys are great. Love your company. Love what you're doing. It's probably the cutting edge work that I've seen. It's needed. This is a future user experience. HTML5, all new stuff, Node.js. This is stuff that's music to our ears. We wish there was more code out there that we could work with. So we'll see you around. You guys in San Francisco? Yeah. San Francisco, obviously doing great stuff. Design is critical, but also the back ends is having that support as well. So we'll be right back with our next guest. Want to thank the...