Nieman Foundation at Harvard
HOME
          
LATEST STORY
Newsonomics: The Daily’s Michael Barbaro on becoming a personality, learning to focus, and Maggie Haberman’s singing
ABOUT                    SUBSCRIBE
April 3, 2017, 11:36 a.m.
Aggregation & Discovery

Filterbubblan is a Swedish effort to give a side-by-side, real-time glance at the country’s filter bubbles

“Very quickly, you learn the worldviews in all these bubbles are very different.”

In the wake of November’s election, the concept of the filter bubble is often discussed as if it’s a uniquely American reflection of a left/right that other countries are somehow immune to.

Not so. In Sweden, concerns about the country’s own potential political filter bubbles helped give birth to Filterbubblan (translation: “The Filter Bubble”), an online tool that gives users a side-by-side, real-time view of the political conversations happening among the country’s political parties. On the left are the liberal parties (represented by red and green); green and blue represent the center parties; and on the right are the more conservative discussions (blue and a darker blue). With a swipe, users can navigate from one feed to the next, simplifying the process of reading about how a topic is discussed in different political circles.

Per Grankvist, an author and columnist for the Swedish newspaper Sydsvenskan, developed the project with Swedish innovation agency Great Works. Grankvist followed the U.S. election from afar and was surprised when Trump won. Why anyone would vote for Trump, he said, “baffled me,” but the widespread support for the then-candidate inspired him to start using Twitter to create lists that could provide more insight into the information sources that fed Trump support. This was invaluable, he said, because it proved that “everyone is voting for a candidate because it feels logical to them. I was trying to emulate that kind of logic. I wanted to recreate that news environment they were in that made the decision to vote for Trump seem logical. That was my starting point.”

Filterbubblan is based around the concept of discrete Twitter lists for different segments of Swedish politics. The project uses Twitter’s suggested user algorithm, which offers users accounts to follow based on the topics and people they’re already interested in. To build each bubble, Grankvist created a new Twitter account and followed the six to 12 accounts suggested for figures central to each bubble, such as party leaders and secretaries. He then kept track of the other accounts that Twitter suggested, adding names until there were around 80 for all three bubbles, each of which represents the feed of a hypothetical person. (The lists don’t include media organizations.) Every tweet that gets pulled in is also analyzed for content and placed in realtime within each feed. Filterbubblan tracks big political topics such as housing, crime, equality, education, and healthcare. These topics, unsurprisingly, are talked about very differently depending on the feed; those on the right, for example, used the recent terror attack in London to talk about the risks of unrestricted immigration.

“Very quickly, you learn the worldviews in all these bubbles are very different. Things are described differently, they refer to different sources, and they often see very different things,” Grankvist said.

Still, politics in Sweden isn’t as hyperpartisan as it can be elsewhere. The Twitter lists that Grankvist built in the U.S. barely overlapped, but he said the filter bubbles in Swedish political discussion are on a continuum, a reflection of how the country’s parliamentary system makes building coalitions more realistic. “In Sweden, we still have conversations and discussion across party lines,” said Grankvist. He pointed out that, in contrast with the U.S., it was a lot more common for Swedes on the left to communicate with and retweet those on the right, and vice versa. “There’s much more of an overlap,” he said.

The project’s resemblance to The Wall Street Journal’s “Red Feed, Blue Feed,” another tool meant to offer users a peek into the feeds of other partisans, isn’t accidental. Filterbubblan was inspired by the Journal’s project, which Grankvist said was “a good starting point” for addressing the issue. But he said that Filterbubblan is a more powerful tool than “Red Feed, Blue Feed,” because it’s based on Twitter’s own algorithm (rather than editorial decisions) and because it places the tweets within the contexts of specific political topics.

Filterbubblan is also similar, at least in its aim, to projects such as Allsides.com and Read Across the Aisle, both of which are designed to help people diversify their news consumption habits and, in turn, gain a deeper understanding of how political discussions look among people they disagree with.

Like all of these projects, Filterbubblan is designed to “put a spotlight on how we talk to each other in the public domain. Everyone is taking a slightly different angle to attack the problem,” Grankvist said.

Photo of the Riksdag Parliament Building by Neil Howard used under a Creative Commons license.

POSTED     April 3, 2017, 11:36 a.m.
SEE MORE ON Aggregation & Discovery
SHARE THIS STORY
   
Show comments  
Show tags
 
Join the 45,000 who get the freshest future-of-journalism news in our daily email.
Newsonomics: The Daily’s Michael Barbaro on becoming a personality, learning to focus, and Maggie Haberman’s singing
“To be a Times reporter is to be in some ways a raconteur, right? A lot of the journalists here are great, great storytellers at a bar…I think The Daily taps into that great oral tradition of journalists, enthusiastically talking about a story in a way they’re excited about, and it gets people excited about it.”
Panoply’s Pinna might just be the first really interesting attempt to get people to pay for podcasts
Plus: 60dB goes to Google, waiting (and waiting) for Apple’s new analytics, and the best podcast-related reads of the past few weeks.
Not a revolution (yet): Data journalism hasn’t changed that much in 4 years, a new paper finds
“Our findings challenge the widespread notion that [data-driven journalism] ‘revolutionizes’ journalism.”