Watch videos with subtitles in your language, upload your videos, create your own subtitles! Click here to learn more on "how to Dotsub"

TEDxSP 2009 - Fernanda Viegas

0 (0 Likes / 0 Dislikes)
So, have been talked a lot about we're on the edge of a revolution a data revolution, an information revolution and if you look, it really looks like we are, right? Because you have public and open data about economy, and everything here, all these data here are true you have a lot of numbers about education, about health about transports. Then I ask, is that the revolution? Maybe is that the revolution? If I show this here to you and ask: "So? Which are you thinking the best car is? Which are you thinking the most economic one? Etc. Which is the best school for your children? How do you understand all this sea of numbers? It's really great to us to have all these information available online but we have to have some kind of translator for us to be able to deal with these information in a human scale. And one of these translators we can use is that piece of technology that's called data visualization which takes this sea of numbers and changes, paints an image allowing you to understand things like: tendencies, anomalies, which is usual, which is unusual in these information you're looking at. This kind of technology, so far were only available to people considered experts. Professional people in the area, scientists serious people, with serious objectives regarding to data. Not to any one. And that is beginning to change. So, with projects like Many Eyes we did at IBM the idea is really to bring technology and to democratize to open, to place this on the web for everyone to be able to use so, Many Eyes is a site which is public, which is for free anyone may take it, to load data, to visualize and to share these visualizations with the rest of the world. So, let me give you an appetizer here this here is a visualization from Many Eyes which some user has made and this here is a visualization from the budget from the American government throughout the years, it comes since 1962 to 2004. And it's an interactive visualization so I can see how the American government is placing money throughout the years, but I can interact I can see, for example, in human resources, which is that big thing there purple has much more categories, is much more detailed than it looks in the first sight. Then I can do the same thing with national defense then I can look at here and... "No, wait a sec, I want only physical resources" then I click here and I see only physical resources or even, inside physical resources, I want to see how they deal with transport air transport, by water, by land energy has a totally different profile. Then you see you are really trying to detail and understand what in this case the government is doing with money and how that changes in the pace of time you see that, in different years, you have a totally different profile about how the government place your money. Now, the cool thing about Many Eyes is that every visualization isn't only a visualization it's an opportunity to talk. So, down here, people came and place n comments questions, hypothesis, people asking: "But wait there. This peak here, I don't understand Why is that so high? What has happened in that time?" Then someone comes and tells: "Ah, it was that scandal, so and so, do you remember?" Because there isn't corruption only here it's everywhere especially with governmental funds and, really, the idea is for a debate to be created which is centered in data, in facts, in official images. So, let's bring it a little bit more to Brazil. Let's see, if we want to view something regarding to Brazil. What would be a system full of complexity? Bovespa. So, this here is a visualization of every companies listed on Bovespa in November, 2009, this year. and, that I'm showing there these data come from Economatica, just for you to know from where this information comes and each company is a little rectangle in this image the size of the rectangle regards to the company's market value so, obviously you'll see that a thing like Petrobras is huge. So, you're seen they're organized in sectors so, there are oil and gas here, there are finances and insurances, which are huge electrical energy, food and drinks, etc. Then, the color of every company regards to the appreciation or depreciation of that company in the last twelve months. So, how much orange the company is better was its performance, more the stock has increased. How much blue it is, worst it was and more the stock fell. So, right now, we can see this one was really bad. This one is Brasil Telecom. There's a difference here, the stock has fell more than 20% in the last twelve months. Petrobras there is keeping the good job. This one here, Hoje X Petroleo is amazingly growing. Then we can do cool things, like: "Now I want to see only a particular sector". Let's see here, telecommunications, so I'm gonna zoom here in telecommunications and I can see all the companies and see that's really a thing kind mixed as some were poor, others were very well for example, this Telenorte here was very well so, whoever has invested here did it great. Then I can see other sectors, for example let's see here, building. Building is great isn't it? Because whoever invested here did it really great. Look at that, there's no blue here. Everybody has grown in building in the last year. So, it's great, if you've invested here, you did it good. But, what is happening? I'm showing to you hundreds of numbers at the same time. A lot of numbers! And, you're able to understand a macro picture from the situation. In the overall, in the macro situation, Bovespa is doing very well. Isn't it? Sure, it has some little issues. And not only you can see the macro, we can also as I've shown to dive and see the little details of each company and to keep this context. That's a thing given to you for free by visualization. Because our eye system is really good to understand very complex patterns. So, everything I've shown so far is visualization from numbers. Sure, when we think on data visualization, ages, we think on numbers. But not ever numbers are the only kind of information we might view. There's another information class that's very important that we may also visualize. That is text. How do we understand a body, a huge collection of text? Does visualization able to help? It does. So, let's visualize another part from Brazilian culture. We've just visualized Bovespa here. Let's visualize a soap opera. Let's visualize "Viver a Vida". A soap opera by Globo. So, what've I done? I've searched like a crazy for the soap operas' script. They don't share it on the Internet if someone knows, come to talk to me later, because I'm very interested on it but what I found was the brief from chapters that thing like: "Ah, on the next week that will happen". So, you don't need to watch the soap opera anymore. So, I've got brief from chapters of three weeks from "Viver a Vida" and I've decided to visualize. Then I've got Many Eyes and I did this visualization. The cool thing is that immediately gave me in one minute the soap opera's social network. And I can see from here that, for example... I don't know who Helena is, but Helena looks to be a central figure. Really central, for her size, that's the frequency she appears on the script. And by the fact she's really on the center there, on the network I can see with who Helena's interacting. We've here Bruno, apparently interacting a lot with Felipe Dora, interacting a lot with Rafaela, etc. This Miguel here, with his connections. So, that's cool, that's a beginning. But that isn't all, because that doesn't tells me what the characters are doing on the soap opera. For that, I'm gonna show you another visualization that really shows the context, which every one is doing and living on the soap opera. Here. So, this one here is what we call Word Tree It takes a text and you perform a search so, I've searched for Helena, because Helena was the central figure which was looking to me on that social network I've searched for Helena and the Word Tree is given to me everything coming after Helena. So, for example, "Helena is", let's see how she is. She knows by Rafaela that Dora is having issues she is embarrassed, she is uncomfortable, she is stunning, etc. Let's see... now I'll look here for Jorge. Jorge... Jorge's also interacting, let's see how Jorge is. Jorge is observing without a job, he is upset, he is bored he is nervous and worried, poor thing. I don't know but he's passing for real trouble on the soap opera. So, the thing happening here is... Ah! there's another one I found very interesting. Isabel, who I found interesting because I can see here things like: "hide the message from Mia", "taunts Renata", "humiliates Mia"... I'm thinking this Isabel is one of the villains in the soap opera, I'm not sure, but looks to me she's doing very shady things. But the thing, what I'm wanting to show is that you can take a lot of texts and with a thing like that, is very easy for you to start to navigate on that text. To understand tendencies, to understand some general things about the text. It's once again given to you a macro view and a well detailed view you're reaching even the sentences on that text. Another thing we can visualize in this way and that's also cool is the Brazilian Constitution. So there's our Constitution. I asked for court, so it's showing to me the different kinds of courts court of law, federal court, which is huge, high court. And then we see here what the court has, the regional... and so on... I'll look for a word here that I've found really cool. Resources. And then it gives a list of the resources specified on the Constitution. Natural, mineral, hydric, public resources, etc. And the human resources, down here very small, they apparently aren't as big as the natural resources of Brazil. One of the things I found interesting, there is natural resources and hydric resources in separated things. And that makes sense, because Brazil is a country so rich in hydric resources that it clearly appears on our Constitution that's a very interesting thing and then you can start to play with that. So, that's the cool thing about visualize numbers and texts, that gives you that... it's how you were wearing a different pair of glasses. Augusto has talked about glasses... it's how you were wearing glasses to look at these things you think you already know in a different way. Now, to finish, the last thing I want to show is that we can also visualize in an interesting way, thoughts. Imagine if we knew what everybody is thinking what everybody is asking to themselves how would that be? There's a way for us to do that. Do you wanna see? Let me get here... Ah, let's go to Google. And you certainly had this experience, getting to Google. I'm on Google and I place a word, for example: "more". And then Google gives to me these suggestions. What are these suggestions? These are the suggestions of the most frequent searches that people did on Google. So, that's what people want to know when they're on Google. So, I'll use that to do a visualization here. Instead to ask for only one search, I'll compare two searches. Let's see how that works. Does he love me? Does he like me yet? But, let's compare this to: "Does she..." Look at this. They also want to know if "does she still in love with me", "does she cheat me". There's more, there's more, the thing gets very interesting. Let's see here what else people want to know... Ah, other thing, Portuguese teachers will be very sad because you see exactly how people use the language, the Portuguese here. So, for example, "por que" unseparated, "'Porque' ele nao ligou?", "'Porque' ele nao me liga?" "'Porque' ela..." and you see, women want to know why he didn't call "Why doesn't he loves me?", "Why doesn't he calls me?" They aren't like that. They want to know why she ignores them. But the thing about "to call" doesn't appear there, what I think, for its own, very interesting. I'm almost finishing here. This one here is also interesting. "How to be popular", "How to be a leader", "smart"... or "How to pretend to be smart?". That is, if you can't be smart, at least you can pretend. Be smart is something like that. There's a sad thing here, "how to..." then I'm comparing this to "how to get..." for us to be a little happier, which is at least the only thing in common here is this "pregnancy" and "how to get pregnant". so, I think there's something good there yet. And then, "I should", "we should...", and the cool thing is that redemption thing to forgive a cheat. You see, there's a theme here. And it's not only things like this, philosophical things only We can ask things... Seu Jorge vs. Marisa Monte, for example. And then we see what people are looking for. I thought that interesting because just in Marisa Monte there's mp3 I don't know exactly what that means but there's a difference. And finally to finish, a thing that I found very, very interesting I put "flirt", and people want to know what "flirt" is I think no one knows anymore, it became obsolete, "to date" everybody understands. Everybody wants to know the meaning too, but not so much as "flirt" and it's just this thing about what it means. Finally, I'll stop here, but I want to remind that, if we're really in a revolution we have to be conscious about this revolution, isn't enough for you to give me a lot of sheets, numbers if I don't know how to read these sheets, if I don't understand what you're trying to say to me with all these data. So, visualization is a way for us to try to start to interact with this richness of information. On Many Eyes for example, we're seen children, students, journalists, activists everybody wanting to use visualization, visualizing any kind of information sometimes personal information and really becoming a consumer much more sophisticated with this kind of information. Understanding what data is, understanding what a graphical presentation is and themselves knowing how to deal with that. It's a question to have your own opinion even when you're relating with data. Thank you.

Video Details

Duration: 18 minutes and 58 seconds
Country: Brazil
License: All rights reserved
Genre: None
Producer: TEDxSP + colmeia
Director: Julio Taubkin
Views: 305
Posted by: tedxsp on Dec 6, 2009

Fernanda Viegas at TEDxSP 2009

Caption and Translate

    Sign In/Register for Dotsub to translate this video.