<p>Been dissing Bradley Wiggin’s sideburns on Twitter? It’s all getting noted by this real-time visualization tool.</p>
If you’ve been following the Olympics, you’ll have your own ideas of how you feel about it. Mostly, from the perspective of the UK, people have been lapping it up—the dramas, the highs, the lows, the strange banter from commentators as they come to the end of yet another 12 hour shift. But of course it’s not just about how the UK’s feeling—the whole world’s been watching it, and you can’t always rely on the impartiality of the British press to divulge what the world really thinks.
For that, we can look to social media. But there’s masses of chatter on the web, too much for any one human to make sense of. For that, we can look to our friends the machines and their buddies, the algorithms. Drew Hemment (FutureEverything), Moritz Stefaner, and Studio NAND have created emoto, a data analysis project that visualizes the emotional response to the 2012 Olympics in real-time using Twitter. They track Olympic-related tweets (only English tweets are used to keep things simple), which they then analyze and visualize according to their various topics and threads of conversation, showing the tone of each one.
Reactions when Bradley Wiggins won gold.
Reactions when Michael Phelps won his 18th, and record-breaking, gold medal.
These are shown as “origami figures” (below) made up of different triangles joined by a circle made up of little dashes. The dashes represent the number of tweets on the topic, which is noted in the center, along with the time the tweets were sent. For the “origami figures,” the top parts show the positive tweets (colored red-orange-yellow), while the bottom figures (colored in hues of blue) show the negative tweets.
The idea is to gauge what people are thinking and feeling—too track their emotions, essentially—and present a sentiment analysis of a series of topics, which can be viewed in the now or historically. The statistics can also be used to create graphs and visualizations (shown on their blog) that can show a topic’s development over time or, for instance, how the British public has been reacting to certain events, like the appearance of Golden post boxes in the home towns of British 2012 gold medalists (below).
(Click to enlarge)
The tweets are collected by connecting to the Twitter streaming API so they can receive tweets in real-time that match a stated query. This is then run through something called “salience analysis,” which sorts out the words that stand for positive or negative emotional responses, scoring the emotion according to how strong it is and giving it a sentiment value.
Once the tweets have been collated and analyzed and assigned a value, they can be turned into the “origami figure” visualizations above and also used for further analysis. The chart below, which isn’t on the site yet, compares the UK versus the rest of the world with respect to different topics. Moritz Stefaner explains: “The vertical axis and color encode the positive or negative sentiments, while the size of a dot represents the number of tweets. Left is always the world, on the right is UK, so if you see a ‘tear’ running to the right bottom corner, you know the UK has tweeted more negatively about that topic.”
But the project doesn’t stop there. Once the games are over all the Twitter interactions will be given physical form as a data art sculpture which will show each day of the Olympics and the ups and downs that came to pass, as seen through the eyes of the tweeting public.
Tweet sentiments and volumes for selected topics. (Click to enlarge.)