Community Post

Bubbles of Choice - NYU Thesis




This project is an interactive video installation that seeks to bring awareness about Filter Bubbles. Those algorithms filter the enormous amount of data available online in order to display only the relevant information to the user. Seeking the right information would take a considerable amount of time with using those pieces of code to help the user. However, if those filters have their benefits, they also tend to bias the options offered to the user. They can ultimately lead to a variety of problems from weakening democracy by manipulating the potential electors to the reinforcement of conspiracy theories just to give a few examples. There are many ways to avoid the negative effects of those algorithms but to know when and how to act against them, the first step is to know their existence. By letting the user experience the world of data and see how it impacts his own filter bubble, I tried to modestly open the discussion about neutrality of the information people have access to.



What is the background of your project? Summarize the research you did, such as your . influences, similar products and projects, and technical research.

Through the year, we had to question ourselves on the usage of technology and the problematics of the data and algorithms. Those topics resonated with me and I was especially interested in the loss of freedom than they could generate. However, those subjects were too vast to be treated in just a project, so I had to reduce the scope. I decided to focus on filter bubbles as it is narrow enough to be treated in a thesis. But I made this choice because those algorithms can have a great impact on the nations and their citizens by being able to shift votes by more than 20% as Jacob N. Shapiro has shown in an academic paper for Princeton University. This is not the only study that shows such results and with the importance of social media and search engines nowadays their impact will probably grow even more. In a year of elections for the United States and just one year away from the presidential election in my own country, France, I thought it was the right topic for my project.

During the IMA Low Res program, I worked on a project to denounce the dark patterns that trick people to do something they don’t want or don’t need and ultimately it was also talking about the loss of decisions made by people. 

In another class, I worked on a project that featured PoseNet and ThreeJs in which I was manipulating some kind of energy spheres. My Thesis has similarities as it requires manipulation of a camera to extract the position of a specific part of the body in order to attach 3D elements to them. For some time, I considered for some time using those coding libraries for my project and also switching the first one for Kinectron in order to use Kinect. However, I decided to use a new software in order to learn the pros and cons of different technologies. Eventually I chose TouchDesigner, as it works natively with the Microsoft device and was created for live interaction.



The project is an interactive video installation that is controlled by people movement. To do so, I used the software TouchDesigner, some Python expressions, and a Kinect for the core of the project. I also used a tiny bit of Adobe Premiere and After Effects to create the intro video. The project has two modes, one is the stand by mode that displays a video of various social media pages and videos that represents the variety of feeds you could find online.  When someone is close enough to the Kinect, the video disappears to let the person enter into the data world that  consists of particles that represent the flow of data. TouchDesigner uses the data position of the spine of the user to attach a 3D element. This one changes color depending on where the person is standing as the screen is divided into 3 zones. Each zone increases the value of one specific color from 0 to 255. If one of the values reaches 255, the 3D shape turns from a sphere to a Cube. If the person moves from one zone to another, the counter on the top right corner shows the calculations as percentage of freedom/objectivity. It acts as a video cue of what is going on. The last piece of interaction is the right hand that can catch the particles in a warp in order to symbolize the collection of information.



I learned a lot technically with this project as I had to learn new software and how to use the Kinect device. The logic of the software has a lot of similarities with coding so for each features that I wanted to implement I wrote how this feature would behave if it was a piece of code of Javascript and then I replace the different functions, if statements, variables and others for loops by the nodes that have similar effects in TouchDesigner and link them with Python expressions. By using a different tool that the one I used through my career and the IMA Low Res program, I noticed the technical limitation of this program and how to optimize the nodes to improve the frames per second and the overall stability. By experiencing the pros and cons, I will be able to choose the best tool for a future project with a bit more certainty. 

If what we learned during this year at the New York University was great, nothing had more importance in the shape of my project than the different user testings. They brought me a lot of great feedback that greatly improved the cohesion of the piece. And for this I will never thank the person who spent time testing my project.

While testing the early version of my project, I came across a few design issues that I had to fix For example, at first particles were made of flying thumbnails of the videos which are displayed in the stand by mode. However, when people started to interact with the project it was really too messy and confusing. So I faded those particles and replaced them with much smaller ones. Also, after testing, I had to color the particles based on the different sections of the screen, in order to give a visual cue to the users. Visually it is less appealing but it allows people to draw connections between the different elements of the project and it was necessary for them to understand how the project was working. After another round of user testing, I decided to add the counter on the right corner to display the percentage of freedom for two reasons. First to link the change of the bubbles to the numbers to reinforce the link with the topic but also to give another visual cue of what is going on. I noticed that while the colors were changing gradually, it was not obvious enough that it was happening all the time. Adding just a single counter was really helpful to link all the features together.

But until the very end of this project, I struggled on how to present the topic of my thesis. To explain why this was the main difficulty of the project, we have to consider the true nature of the Filter Bubbles. They are algorithms that are hidden, no one sees those algorithms in action, no one chooses their parameters except the one who created them and maintains them. Because they are not aware of them, people cannot talk about the consequences of those filters and cannot act against them if they think they are bad for themselves. That’s why,, it was complicated to present this project without any explanations, people could have enjoyed the interactions but would not have understood the topic of the project, and therefore it would have become meaningless. And even if the different user testings showed that the different additions were going in the right direction, it was never enough, not clear enough. So, I had to start the projection with a question to tell the user what this project was about. 

I may have been available to vehiculate this information by making the project less abstract by using clear reference to a specific topic such as the American election by displaying democrats and republicans news and videos. However, if it may have worked in the US it wouldn’t have been relevant in other countries. And no topics would have been universal enough, as people are so unique and there is an infinity of information available online, it wouldn’t have made sense to make a project displaying only a few types and hope it will be sufficient. That is why, my project has a more abstract look in the end. I hope that my project will be understood enough to start the conversation about Filter Bubbles and that people will regain a tiny bit of their freedom of choice and objectivity.



Academic Papers:

Quentin André, Ziv Carmon, Klaus Wertenbroch, Alia Crum, Douglas Frank, William Goldstein, Joel Huber, Leaf van Boven, Bernd Weber & Haiyang Yang. “Consumer Choice and Autonomy in the Age of Artificial Intelligence and Big Data”, Dec 19, 2017

Henrik SkaugSætra.”The tyranny of perceived opinion: Freedom and information in the era of big data”, Østfold University College, Jul 5, 2019 Czopek. “The phenomenon of the filter bubble as a threat to freedom on the Web”, International Journal of Pedagogy, Vol 7, p11-15, Oct 12, 2020 

Engin Bozdag & Job Timmermans. “Values in the filter bubble Ethics of Personalization Algorithms in Cloud Computing”, Delft University of Technology, p7-15, Sep 11, 2011

Kim Mortimer. “Understanding Conspiracy Online: Social Media and the Spread of Suspicious Thinking”, Dalhousie University, 2017


Engin Bozdag & Jeroen van den Hoven. “Breaking the filter bubble: democracy and design”, Dec 18, 2015 “How Filter Bubbles Distort Reality: Everything You Need to Know” 



Eli Pariser. “The Filter Bubble: How the New Personalized Web Is Changing What We Read and How We Think”, Penguin Books, Apr 24, 2012
Nicholas Carr. ”The Shallows: What the Internet Is Doing to Our Brains”, W. W. Norton & Company, Jun 6, 2011
Evgeny Morozov. “The Net Delusion: The Dark Side of Internet Freedom”, PublicAffairs, Feb 28, 2012


I wanted to thank Matthew Ragan, Elburz & all the TouchDesigner community for their tutoriels and help on the forum of Derivative.

Finally, to keep the best for the end, I would like to thank all the IMA Low Res professors and all my classmates who made this thesis possible. I wouldn’t have been able to do it without you all.



Youtube Video:

Andymerv. “The Funniest Interview You Will Ever See!”, including footage of the TV show Boemerang, Nov 8, 2008

Edigeevich. "Drunk Boris Yeltsin", Oct 10, 2013

Franmunication. "Sheikh Bandar al Khaibari-He rejects that Earth revolves around the Sun", original footage from Al Arabiya, Feb 18, 2015

Vox. "ISIS videos are sickening. They’re also really effective", Jul 7, 2015Faizal Saban (Faizter). “El Risitas - Funniest Laugh Ever!!”, Apr 15, 2016

Red Bull. "Skysurfing in an Active Thunderstorm | Storm's Edge", Jul 21, 2016Radio-Canada Info. "Revivez le 11 Septembre en 2 minutes", Sep 10, 2016

Pr. Rachid Eljay chaine officielle 1. "SOIS PAS TRISTE, RELÈVE LA TÊTE ET GARDE LE MORAL!", Feb 19, 2017

Travel Thirsty. “Japanese Street Food - $600 GIANT RAINBOW LOBSTER Sashimi Japan Seafood”, Apr 20, 2018

The Guardian. "Flat Earth rising: meet the people casting aside 2,500 years of science", Feb 5, 2019ABC7. “Black Lives Matter protesters smash window of CHP cruisers near downtown LA I ABC7”, May 28, 2020

CNBC Television. "President Donald Trump releases video in response to the Capitol riot", Jan 8, 2021

VICE News, "Inside the Capitol Hill Riots", Jan 10, 2021

BBC News. "President Biden inauguration speech in full - BBC News", Jan 21, 2021

BBC News. “Anti-Asian violence in the US: 'He slashed me from cheek-to-cheek'”, Mar 20, 2021

Vatican News. "03 April 2021, The Easter Vigil - Homily, Pope Francis", Apr 4, 2021Video:Pixabay, "Little Kitten Playing His Toy Mouse", Feb15, 2017, Pexels

Evgenia Kirpichnikova, "Woman Dancing", Aug 31, 2019, Pexels

Kelly Lacy, "Police Officers In Line To Keep Order In The Protest", Jun 12, 2020, Pexels

Pavel Danilyuk, "Man Climbing a Wall", Apr 21, 2021, Pexels



Barack Obama, “”



Donald Trump, “@realDonaldTrump.” Account suspended