In the Filter Bubble of Raw Data

The Netherlands are unsafe.

The Netherlands are being islamized.

The Refugees are threatening our well-fare state.

– Prior to the European National election marathon (The Netherlands, France, Germany), many so-called “gut feelings”, created by rumors, drastic news reports, and good old xenophobic sentiment are the sparking hope of the New Right – populist political parties on the rise, yelling hateful slogans as the simplifying solution to an ever more complex world. is one of several initiatives to fix the public debate’s distortion by these supposedly fictional gut feelings with the means of illustrating so-called facts with simple graphs that should debunk the myths as fake. However, when taking a closer look at the illustrations, it quickly becomes apparent that the handling and presentation of assumed raw data is tricky, no matter on which political side the author may stand: One of the graphs is shown to fight the myth of the world being unsafe.Screen Shot 2017-03-31 at 22.05.44

On a timeline beginning in the 70s and going until today, the data visualization shows how between 1975 and 1995 the number of terror attacks in Western Europe (which is apparently “the world”) were consistently higher than in the years after 2000. From that is concluded that, contrary to the gut feeling that the world is unsafe, the world actually became safer in the last few decades.
The problem with this “fact” is that it is in no way less relative and therefore subjective than the statement that the world is unsafe. Reading the latter as “terror attacks currently make the world unsafe”, it could be easily backed with the same graph if one was to simply start the timeline at 2005 (just like the original authors decided to start it in 1970).

To address this difficult relationship between data and opinion and the resulting general skepticism towards top-down provided data, Dutch media lab SETUP in collaboration with research platforms The Datafied Society and [urban interfaces] invited artists, researchers, and students to a one-day hackathon with the mission to design an installation in public space that allows its audience to interact with data deemed relevant to the upcoming election and therefore make it more tangible for the everyday Dutch citizen that gets to give their vote in the 2017 election.
The hackathon took place on 8th March (one week prior to the elections) at the SETUP office in Utrecht. The morning was used for an extensive discussion between the participants on issues of data politics and set the stage for an afternoon of creating two installation prototypes that were finally presented to a semi-public audience.

The morning discussion started from’s example and uncovered several data-related issues relevant to the upcoming election: Participants agreed on the basic problem that the individual voter lacks a feeling of personal responsibility, stemming both in the feeling that one vote cannot make a difference and that voting is more a representation of loyalty to the respective group the voter is part of, than expressing personal beliefs. That second issue is highly connected to the much cited phenomenon of the information Filter Bubble, people are supposedly willfully embedded in, where populists address hateful sentiments, conspiracy theorists work on the undermining of factual storytelling, and teams of voters exert peer pressure to keep all their sheep in the herd of streamlined political opinion.
Connecting this discussion with the proceedings of the critical media lab IMPAKT event “Filter Bubbles and Fake News” on 12th March, it becomes apparent that the hackathon participants tapped into a more complex media phenomenon than simply an online populist circle jerk: As media and culture studies researcher Dan Hassler-Forest laid out, there are at least three different ways in which the Filter Bubble mechanism affects and shapes our intake of information, and we all fall victim to this.
Firstly, there is the most-commonly addressed algorithm-based selection of news and information items that fit our pre-existing and pre-recorded (political) preferences: If I heavily clicked news items on immigration in the past, any selection algorithm like Facebook’s or Twitter’s will be more likely to show me the latest news on migrants and how they are dealt with.
This feeding with more of the same is secondly further boosted by the commercialization of mainstream media and its alignment to the assumptions of Attention Economics: On the one hand by the phenomenon that a “successful” news item on a certain issue will result in the production of more news items on similar issues and therefore an overrepresentation and over-awareness in the public sphere, and on the other hand that more disruptive (e.g. outrageous) news items generate more clicks/sales and therefore more (ad-)revenue, resulting in lower interest in balanced and well-investigated coverage.
These principles are successfully adopted by populist politicians, who managed to acquire extensive media coverage by performing outrageous utterances and therefore gaining mainstream fame without any political labor or achievements on their record (see e.g. the essentially free press coverage of Trump’s election campaign).
Finally, the Filter Bubble by means of the equally heavily cited phenomenon of Fake News operates as a strong driver to the dissolution of the border between fiction and facts. As renowned podcast host Benjamen Walker establishes in his interview with MIT Center for Civic Media director Evan Zuckerman Fake News in the form of Disinformation undermine the public belief in the very existence of objective facts:

“You simply pump out as much dezinformatsiya as possible, until people start doubting. Doubting what is under investigation, doubting what was in the e-mails, doubting what was going on in the hotel room with the Russian prostitutes. Doubting what is true, and doubting what is false. – And the people who benefit from this sort of doubt that’s set on by dezinformatsiya are strong charismatic leaders.”

Furthermore, the Filter Bubble functions as verification loop for its own fiction and the opinion leaders that reinforce it: When a myth, spread by one of the Filter Bubble “news” sources (e.g. the infamous Breitbart News), is repeated by an official figure (like President Trump), the loop justifies it for the filter bubble audience with the impression that the opinion leader “finally speaks out loud what all the others tried to hide from us”.
With this focus on the general manipulation and manipulability of narratives derived from assumed “raw data” (Karin van Es: “There is no raw data.”), the hackathon participants shifted away from the originally intended goal of creating an installation that allows Dutch citizens to fact-check their gut feelings, and in the following prototyping session aimed at debunking the data myth and finding alternatives:

The first design prototype started off from the idea that the same seemingly neutral data can be interpreted differently by different people and adopted and adapted for different ends by telling different data stories from the same data base: The team created the playful installation Realiteitmachine (a word joke combining the Dutch words for time machine and reality, playing on the idea of an imaginary machine that can be (ab-)used to alter reality), a device that would ask the player to verify any political claim with potentially contradicting data stories from one and the same data pool.
At the beginning of a play session the user sits down at the desk that comprises the Realiteitmachine and receives a Mission Card in a sealed envelope. This card would give a political statement (e.g. “The Netherlands should close the borders to all new immigrants.”; all statements derived from the Dutch election advice webpage to which the player’s position has to be determined by rolling a die (“in favor” or “against”).

The player then has to put on the rose-tinted Data Goggles (plastic toy goggles) and fill the machine’s Data Storyboard, consisting in a large free spot for statistics and graphs and smaller free spots for pictures and newspaper headlines, with material from the four categories of data on the Data Pool Flipchart: immigration, national safety, inequality, economy. Eventually the player would present the constructed data story that supports the random position on the random claim and could share it for comparison and discussion – physically in an adjacent gallery, and virtually on social media like Twitter, complemented with a hashtag stating the player’s own position on the claim (#iftheywerethem vs. #iftheywereus).
The audience reactions to the prototype’s presentation were mixed: While some noted that the Realiteitmachine succeeds to make the deconstruction of the composition of data stories accessible in a playful manner and raises awareness for the need of data literacy, others found the playful approach towards partly hateful and discriminating political claims scary and cynical and demanded a toning down of the installation’s harshness. Also, it was noted that instead of motivating higher personal involvement with the construction of data stories, the installation actually promotes a fatalistic view on data, in which it is seen as completely malleable and therefore useless in public debate.
As part of the installation’s creator team, while unable to properly judge the installation’s overall value and success, I would like to respond to the criticism that the Realiteitsmachine in my understanding only makes sense in this highly cynical way, as only then it can create a disruption strong enough to actually affect the user: By playfully constructing a potentially hateful data story to back up a bigot’s claim, simply because it was the mission of the game, the player becomes complicit in the everyday deception of the public by manipulating and creatively adapting seemingly raw data and therefore has to deal with it as a personal issue and cannot dismiss it as something only “bad” people or populists do. Consequentially the installation indeed does claim that data is not the solution to questions of public discourse, as we are too good at manipulation and too susceptible for its colorful simplifications.

The question that is left open then, though, is what else than data could help us bridge the gap between gut feelings and real world situation? The second creator team delivered an installation prototype that is supposed to deal with this missing link by concreteness rather than simplification:
The Realiteit Bureau is based on an earlier installation of the group, meant to connect real people from different backgrounds and therefore debunk myths and stereotypes held in society about these backgrounds by establishing personal connection. In this application, directed at the stereotypes potential populist party voters in The Netherlands have towards immigrants, a mobile white wooden wall holds two smartphones, one of which is playing a prerecorded video of the member of one of the groups asking a personal question, while the other is set up to record the other group’s physically present member’s answer. Afterwards, both videos are made accessible on the installation’s webpage with the intended effect of members of both groups feeling personally involved and connected and hereby helping them to overcome their stereotypes that are believed to ultimately be based in a fear of the unknown.

Realiteit Bureau1
While in the audience reaction toward its presentation, it was noted that the installation itself could reinforce stereotypes by assuming that Dutch potential populist voters and immigrants are definable and therefore somewhat homogeneous groups, it was also praised for expanding the notion of data from merely quantitative abstractions to more qualitative artifacts such as video recordings.
Furthermore, the Realiteit Bureau is not limited to establishing Dutch-immigrant dialogue, but can be used to set up pseudo face-to-face and therefore less threatening contact between people in general and can therefore be employed to address various issues of public debate.

In conclusion, both installation prototypes offer ways out of the vicious cyclic mechanisms of the Filter Bubble: The overall skepticism towards data stories, generated by the Realiteitmachine, can bring the simplifying narratives told by populist leaders and raging “news” reports into question, when the consumer understands how easy it is to find fitting data to back up nearly any claim. In addition, the Realiteit Bureau explicitly breaks the algorithmic cycle of more and evermore of the same by putting real people with agency in the foreground and therefore offering an alternative to patronizing and simplifying abstractions.
Both installations do not vilify data in a general way, but demand reading more source and investigating where they come from and go beyond that, by becoming personally responsible, talking to people on the horizontal level rather than only listening to those in power.
Or, to sum it up with Edward Snowden: “The answer to bad speech is not censorship” and therefore control, but “the answer to bad speech is more speech”, hence activism.


2 thoughts on “In the Filter Bubble of Raw Data”

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s