About Carol Rozendo

I am a web designer from Brazil who's been working in the news industry for the last 6 years, creating interactive narrative and data visualization. I am a visiting student at the MIT Media Lab.

Social Reality Check: is our social network an ideology bubble?

Disclaimer: As a designer, I prefer to take on a personal point of view instead of trying to turn myself into a good reporter. There’s also the fact that I am not an extraordinary English speaker (and too perfectionist to think this is ok). So bear with me. ūüėČ

Last week, Miguel and I were talking about the conservative wave that has been sweeping the social networks in Brazil. Our subject, at the time, was the recent implementation of a network of cycle paths in the capital of S√£o Paulo, the largest and most congested city in the country.

S√£o Paulo has way too many cars. For a population of around 12 million, the state capital registered 5.4 million cars – roughly one for every two people, according to the state’s Department of Motor Vehicles (Detran). The metro network in the city is burdened, overcrowded, and does not reach the entire urban area. Buses are also crowded and subject to traffic conditions.

In June 2013 people marched in the city, protesting against the increase in bus and subway fares. When cases of police violence against demonstrators became public, the popular movement expanded and millions of people took to the streets of the whole country.

After these events, the mayor of S√£o Paulo Fernando Haddad attempted to reduce traffic chaos through unpopular actions such as creating exclusive bus lanes, improving fleets and remodeling public roads, culminating with the implementation of a network of cycle paths in 2014. At this time, a war between the drivers culture and the bike culture¬†drew up very clearly on social networks, and soon transformed into a textual war between¬†individualists and collectivists. The mayor received a barrage of criticism from angry drivers who complained about the reduced space in the streets for their cars. Most pedestrians and cyclists, however, approved, and Haddad’s popularity grew among young people.

The mayor of S√£o Paulo, Fernando Haddad, eventually uses his bike to commute

The mayor of S√£o Paulo, Fernando Haddad, eventually uses his bike to commute

My own timeline of Facebook, however, seemed almost immune to that criticism against the cycle paths. Messages with very conservative content rarely appear in my Facebook feed, and I was only aware of them because of comments made by outraged friends or articles in magazines and newspapers.

From this fact came the idea to address the ideological bubbles on social networks. Do our private feeds on social networks correspond to general opinion in the real world?

With the recent protests in Brazil against the rule of President Dilma Rousseff, we thought that the war we spoke of would certainly intensify. It would be interesting to compare the Facebook public news feed to the private feeds of specific users, trying to create a visualization of the results.

Using Facebook API, I scanned the public stream for posts in Portuguese containing the name of the president, and stored dozens of files like this one, so that I could analyse the results. Unfortunately, however, I could not code the search in private news feeds in time to follow the course of events in social networks, so there was not enough data for a good comparison.

But here is a word cloud generated from public posts collected during the protests:
Word Cloud - Public stream (Dilma Rousseff)
And here, a cloud generated from newsfeeds of five people who considered themselves contrary to the protests of March 15.
Word Cloud - Private stream (Dilma Rousseff)
An interesting thing to notice is the frequency of the words ‘direita’ (rightist) and ‘ditadura’ (dictatorship) in the private cloud. Among people contrary to the recent protests, the general opinion is that the participants of such demonstrations are people from the white middle class, mostly individualistic, and supporters of the dictatorship. It may be also significant that in the public timeline the word ‘protest’ is more frequent than ‘demonstration’, while in private timelines this relationship is reversed.
Unfortunately, we could not get to the in-depth comparative analysis we intended this week, but with more time I think I would try not the word cloud format, but the word network format, as in Textexture tool website.
Posted in All

Exhibition: Drawing apart

Since I was not confident about writing a lengthy article in a language that is not my native one, I decided to explore the format and presentation.

I am not at all a professional photographer, as you will notice, but as I visited the very small exhibition I wondered how I could make the format of the report express a bit of the experience of being there. So I chose to take close-up, non revealing pictures of the works exhibited, and to fragment those pictures as the artist chose to fragment her scale models. If you click on a block, it rebuilds the correspondent image.


I cheated slightly, to be honest. My camera battery died right after I went to the exhibition, so I had to buy a charger before I could finish the assignment.

Click here to see the result

Carol’s Media Diary

I installed RescueTime on Chrome and on my Android device. One of the interesting things I noticed is that, by default, RescueTime considers news reading as unproductive time. You can change that, but I started thinking about what productivity really means, as the concept may vary greatly depending on you field of work. I also had to make some corrections, for example, in cases where I watched videos on YouTube, but they were instructional and couldn’t be accounted in the entertainment category.

I had a look at the graphs provided by the application and took some notes. It may seem a little obvious, but it was interesting to see that most of my media consumption happened at the weekends. Also, I realized I haven’t been reading much news, for many of the articles I read in that category were actually opinion articles on culture, music or art.

So yes, I noticed is that I have a “high entertainment” media diet. But the information may be a bit misleading because I listen to music on Spotify or YouTube while working. For that reason I chose to restrict the color palette so that periods of media consumption higher than 3 hours wouldn’t impact the visualization too much.

I also grouped the media consumed by format: video, audio or text & other. This was not very accurate, for I simply tried to remember what was the media format for that link. But it is still interesting to notice that a big part of the media I consume is in text format. Photo galleries and some applications¬†that were hard to define are grouped in the “text & other” category, but they represented a very small part in the whole.


Posted in All