Tag Archives: data visualization

Breadth Portfolio: Part 3 – Data Visualization

The last part of the series with parts of the report I made for the Multimedia Journalism module, this time about data visualization.

Experiments in Data Visualization

One of the fields I’ve been interested the most is in data journalism, and the visual representation of information. It takes two seemingly opposite mindsets to work with data: one of a statistician and the other of a designer, the analytical and the creative side by side. But I found that data, to be interesting to the audience must have one or both of these features, besides being accurate and relevant: it has to be visually compelling and/or interactive. Once again, technology comes to the rescue, and at the same time can lead us to disaster. The huge amount of tools available to organize and present data relies in different coding languages, mostly Javascript and Flash, and if we are to use live data we must know how to use APIs to direct content into our application. And even if we find a software that does it all for us, we need to know which story we’re telling.

I did some research about newspaper brand values and online traffic of Portuguese news websites, and I came across with the monthly traffic report for all of them. My goal was to understand the relative and proportional position of each one, regarding visits, page views, and how those two values relate to each other. The data I got also has portals, specialized websites, and entertainment magazines so it has a broad range of themes (all charts are available live here – http://is.gd/aZLXs)

First of all I wanted to have a general overview of the size of each one when it comes to space within the Portuguese online universe, so I went for a tree map. The view wasn’t clear enough, so I tried a different approach, using a bubble chart, highlighting just the focus of my research, which were specifically newspapers. It was a better option, and I could have done things even more interesting if I added the paper circulation data. One of the conclusions that I would have found was that best selling newspapers don’t necessarily do equally well online.

But the real risk when connecting data is to draw wrong conclusions from fact. So these are the most successful websites, looking at these data sets right? Well, maybe not. What if we try to understand who are better at engaging the audience and get more page views per visit?

The idea that I got is that we can use data visualization to find a perspective on the subject and that will lead us to a better understanding of what were just numbers on a paper.