 We have two microphones, so please just raise your hand if you want to ask a question or want to dig deeper into something one two Thank you. Hi Nicholas. I'm a question to Georgia. I See the point of choosing a point of view But with which kind of point of view did you do you not choose? Sure, this is a good question. Of course as I was really trying to introduce There's many many many ways a story can be told. I mean, there's many points of view on that We try to I mean just to be honest We have really really a small amount of time because we do the data analysis and the data visualization Just four days and we really try to do our best to find how the most interesting thing that appears From the data being something that we didn't expect it being something that can open to further exploration So that just is a starting point for readers to understand the topic and you really have every time try to Correlate different kind of information which I know your point is not the only point of view But it's part of the job We know that we are missing thousands of point of view But this is a story that we tell in a Sunday culture of supplement of a newspaper So it's really really a story that we we choose our Point of view to tell the story through so I totally get your point We miss lots of point of view every time and you know every time that we find an interesting data set and then we fit The data in print we always say oh, and this is so amazing We have to do it interactive so that people can explore all of the layers But most of the time we're not able to do that but yeah, I see your point Yes, please Hi, yeah, that was my questions if you have made any interactive Visualizations and yeah, and what do you think of that? Yeah, so we did I mean today as I talk and then you can all As actually that we do a lot of different kind of information design work So we build also interactive visualization interfaces and after and that's it today. I Show I choose to show this kind of work Which is the pure data visualization work with the data and the most exploratory work when it comes to interactive view Of course as the possibility to let Readers and so users find their stories through data So it's more that you build a tool a sort of infrastructure when you present the data And you give the possibility to let the readers and the user interact in their own way finding their own path and really digging into the thing that they Are willing to explore of course when you are in static data visualization You have to pick a point of view you have to let something Over of course, so I think that of course interactive data visualization also in the visualization that can really Let users to explorers and to find their path is more Exploratory tools I would call lots of interactive data visualization as tool to explore rather than pieces to read So this is my difference. I don't know so and for us everything at the at the New York Times R&D lab that we try to build We always try and build something that you can play with something that we can touch and we can test out in the real world and see How effective a visualization is or how effective a prototype it is so it's it's really important to us to make things that Are instantiated in one way or another so for the globe visualization that I showed The live version of that means you can spin it around and zoom in and out and really try and Get in amongst the data Yes So I have a question for you Mike and the maps over Afghanistan that you Did what I used afterwards for the like to raise the security in Afghanistan for the military or anything? Yeah, so immediately after we made the maps we One of my colleagues got a call from a think tank in Washington and they use them For the kind of arguments they were making for how they should proceed in Afghanistan We ended up doing a bunch of academic work around it. So there's a paper and there's a book coming out all about how to model the Conflict dynamics, so it's gone in a number of different directions Can I have a follow-up question on that Mike? What's the biggest aha moment you've experienced you or your readers when Visualizing big data In general so for the for the Afghanistan map it was It was definitely the relationship between the conflict and the road that the NATO forces were building around the The edge of the country so you can really see you don't even need to plot the road You can see where it is from the ring of activity and you can see the The most activity taking place where the edges of the road are being built So and with every kind of visualization we always get some kind of correlation that suggests more questions So all of the things we do Are always to kind of to get in there to start asking more subtle questions Mike for you The data about the users of the New York time page. Where did it actually go? Did it go into a boardroom? Where management would be thinking about solving the New York Times dropping leader in readership or what were the What's the audience? so The audience for that is definitely to go into the rest of the building and should try and persuade The the organization to maybe think a little differently that particular piece of work started getting implemented a month or so ago so a Good a good success from a project from the lab is when it gets to go and cause ripples in the in the organization And this one was a particularly good one for that So yeah, hopefully it'll be implemented and we'll start seeing what how people use the site a bit more and less about What they're reading? Hi, my name is Julia and my question is for both of you And I wondering how do you build your interactive designs in what programming and stuff like that? Well It depends we at Acura. We don't have a fixed technology that we use it depends from the aim I think that our developers right now mainly work on D3 and Customize their programming, but I have to admit I have no idea because I'm not a coder So as I was saying before for us is very very important to have build a team because I'm in charge of the representation of the information And every time I can talk with coders that can really do their job And also advise me which could be best ways to maybe Dig into the data visually speaking. So but right now I think that mainly we use D3 Yeah, so we've Over the last few years. We've been exploring lots of different technologies. So I think in the the talk you saw some Python and some GG plot in our and then the more recent stuff is all JavaScript. So the two the two visualizations that Nick made are both in canvas And he's but he's not a good example He's a bit of a crazy whiz kid and uses the graphics the GPU to do a lot of that work Other stuff that we've done. We use a lot of D3 and JavaScript based things It's really good for being able to build a really complicated visualization and get it in front of a lot of people quickly a lot better than The just building little the kind of more scientific plots, which you have to embed into a PDF and then persuade people to open Just clicking through on to a JavaScript based interactive is so much easier stir question There's a guy behind the screen Way back. Hi, my name is Ola This one's for Mike This team that works with data visualization based on big data Has it ever come look to something like a grounded theory, which is a social science methodology for abductive theory generation Not that I know I'm afraid what's the name of the theory grounded a theory It's grounded in data trying to find the variances and builds theory from the patterns that are Identified rather than the hypotheses. There's formed ahead of time, right? So I'm without much discipline That's definitely the way that we work, but I would love to know more about the The discipline and the the written-down theory would be fantastic We've all got really disparate backgrounds in the lab. So Theories like that don't show up in an engineering course. Unfortunately. So I'd love to know more In the discipline is a is a hard thing to accomplish Okay, I do have a final question and when you do this, it feels like it's it's lots of Manual work you go through your data You decide what to scrap and what to not then and your work is almost like artwork I mean, it's lots of hours goes into it. Are there any pieces you could automate? To make them quicker or more more direct or is it is it always this handy work that needs to be done? collecting and and selecting and So for us where we work a lot every project gives rise to new tools And then we try and keep those tools in into the next project and so We've we're building up a really interesting tool set that's all on github if anyone's interested It allows us to do these these slightly more impressive engineering feats So to get all of the click all of the page views on the New York Times into one browser window has built up on lots of Other tools that we've built up over the last I guess the last few years In our case for this project everything is very manual and very customized to the piece that we want to build So we basically work with excel and illustrator and sometimes you can of course Automate the way that the most part of information are represented But everything that is really refined in illustrator trying really to be very precise So it's really passing from excel to illustrator. Okay So I would like to say thank you for making us understand complex things better. Thank you very much for great presentations and