Pop Your
Bubble

A news application that makes the invisible visible — exposing algorithmic filter bubbles and offering an alternative perspective on every story.
ROLE

Solo UX/UI Designer

Type

Vordiplom/Pre-Diploma

CONTEXT

10 weeks (2022)

TOOLS

Figma

PROBLEM 01/06

News apps personalise aggressively but never show users the filter shaping what they read.

News apps personalise aggressively but never show users the filter shaping what they read.

News apps personalise aggressively but never show users the filter shaping what they read.

INVISIBLE CURATION

What is removed from a feed is never surfaced users cannot know what they are not seeing.

LOCKED PROFILES

Algorithmic identities accumulate silently and are impossible for the user to inspect or reset.

NO COUNTER-VIEW

Every news app delivers one version of a story there is no built-in mechanism to read it otherwise.

NEGATIVITY DEFAULT

Mainstream feeds optimise for engagement; positive and constructive stories are systematically deprioritised.

DESIGN APPROACH 02/06

One app.
Two persistent views.

One app.
Two persistent views.

One app.
Two persistent views.

Rather than removing personalisation, Pop Your Bubble keeps it — and surfaces it. Every article comes with a toggle: YOUR VIEW (your algorithmic profile, black UI) and OTHER VIEW (a second profile you configure, white UI). The colour assignment is intentional: black reads as fixed and opaque, white as open and configurable. Neither is framed as neutral.

01

BLACK UI

YOUR VIEW

KEY FUNCTIONS

Generated from behavioural data — read-only

Personality profile shown via Five-Factor Model

Updates automatically as reading patterns shift

02

WHITE UI

OTHER VIEW

KEY FUNCTIONS

Up to 8 manually configured personality profiles

Or connect directly to a friend's real profile

Renders the same feed through a different filter

03

DISCOVERY

EXPLORE & HAPPY PAGE

KEY FUNCTIONS

Explore: articles outside your current interest profile

Percentage indicator shows distance from your bubble per article

Happy Page: positive global news, interest-balanced

DESIGN 03/06

Building your bubble without filling in a form.

Building your bubble without filling in a form.

Building your bubble without filling in a form.

Registration skips topic checkboxes entirely. Instead users choose between image pairs in under 3 seconds — time pressure forces intuitive choices the algorithm can use reliably. Hesitation is registered as a signal too. The output is a personality profile across five dimensions, not a list of interests.

Image pairs,
not topic checkboxes

Forces intuitive choices; users cannot construct a flattering false profile.

3-second time limit

Decision delay is itself a signal; hesitation reveals genuine interests the algorithm can weight.

Five-Factor Personality Model output

Maps to established psychological dimensions so the profile display is legible, not arbitrary.

Up to 8 configurable Other View profiles

Sliders adjust each personality dimension independently; profiles are saved and switchable at any time.

Connect to a friend's real profile

The most direct form of perspective shift: the feed is literally someone else's, not a hypothetical approximation.

Four reading modes — Global, Local, Sublocal, Happy Page.

Four reading modes — Global, Local, Sublocal, Happy Page.

Four reading modes — Global, Local, Sublocal, Happy Page.

Every section carries the YOUR VIEW / OTHER VIEW toggle. Global and Local follow standard news categories. Sublocal surfaces stories within a 60 km radius and pairs them with an equivalent event elsewhere in the world. The Happy Page is positive news only, balanced across interest areas regardless of the user's profile.

Black for YOUR VIEW,
white for OTHER VIEW

Neither framed as neutral; the colour inversion is the conceptual argument, not decoration.

Same topic,
different editorial framing

Both articles address the same subject; only the perspective and emphasis shifts, not the subject matter.

Category tabs persist
across both views

Politic, Economics, Opinion remain consistent; only the algorithmic filter changes, not the navigation taxonomy.

WHY SUBLOCAL WORKS THIS WAY

YOUR VIEW shows stories within 60 km — local relevance is preserved. The OTHER VIEW mirrors the same event type elsewhere in the world. Users see that their situation has a parallel somewhere else, without losing the local anchor.

THE ALTERNATIVE REJECTED

Showing an unrelated global story in the Other View. Rejected because it loses the structural parallel — matching event type to event type is more legible and more meaningful than simply swapping geographic scope.

DESIGN System 04/06

Helvetica Neue
2-colour system Unfarbigkeit

The palette is reduced to two colours — black and white — named Unfarbigkeit (un-colour) in the project documentation.
Black carries YOUR VIEW screens: fixed, generated by behaviour, uneditable by the user.
White carries OTHER VIEW screens: open, configurable, a blank surface for a different perspective.
An optional setting lets users switch all article images to greyscale — surfacing how much colour shapes emotional reception of the same information.

Reflection 05/06
WHAT WORKED

The two-colour architecture communicated the concept without needing explanation — the black/white split was immediately readable as two distinct perspectives. The gamified onboardingsolved the cold-start problem neatly: no topic lists, no blank slate, no self-selection bias. The Sublocal pairing logic — local story matched to a geographically equivalent story abroad — was the most original structural idea and differentiates the app clearly from existing news products.

WHAT I WOULD DO DIFFERENTLY

The Other View slider configuration assumes users understand what adjusting "conservative vs. curious" produces in a feed — that mapping is not self-evident and would need clearer onboarding. I would also run user tests on the colour inversion earlier in the process; the black YOUR VIEW risks reading as negative rather than simply fixed. The Happy Pageneeds sharper editorial criteria — "positive news" is too broad a brief to hand to a content team in production.

resources 06/06

Pariser, E. (2012) Filter Bubble: Wie wir im Internet entmündigt werden (E-book). Munich: Carl Hanser Verlag. ISBN 978-3-446-43116-4.