JM
Jan MalyCollective decision-makingmy interview
"If people can't see where feedback goes, they stop giving it."
Changed: the "sent to the city council" line on the results screen.
Oct 2024 – Jun 2025
Design ResearchInteraction DesignPhysical ComputingPublic Installation
TU Wien · Design Thinking I & II · with aspern.mobil LAB
Consensus Wheel
Cities say they want resident input. What they offer is surveys and town halls, which reach the same handful of people every time. So we built a wheel you spin on the street.
My role
Interaction design & project planning
Duration
2 semesters
Team
5 students
Ended as
Working prototype, exhibited
The 30-second version
The problem
Cities want resident input, but surveys and town halls reach the same few voices.
What we made
A hand-spun street wheel wired to a colour sensor. Spin, answer, see the neighbourhood's votes.
My part
Interaction design, keeping the build technically feasible, and project planning across both semesters.
What it proved
Five testers, zero instructions, five successful spins.
Oct 2024
Cultural probes in the field
Nov 2024
Expert interviews & literature
Dec 2024
Design workshop & game
Mar 2025
Sketches, wireframe test
Apr 2025
Cardboard mockup test
May 2025
Hardware build, final test
Jun 2025
Public exhibition
●Try it first
Spin it yourself
A working recreation of the real flow. Interaction is real; vote counts are illustrative.
01The brief
One word: consensus
TU Wien gave us Seestadt Aspern, a district being built from scratch on the edge of Vienna, and one theme. Everything else was ours to figure out.
The city genuinely wants resident input. The problem is the format: town halls and paper surveys.
Group decisions are slow
A few voices dominate
Most people never participate
How might feedback feel less like paperwork and more like play?
02Team & my role
Five people, one table
Tightly knit team, long shared sessions. Named roles below, but most of us touched most things.
L
Loriane Unterweger
Project lead
R
Rita Kerkovits
Presentation & documentation
G
Gledisa Balliu me
Interaction, feasibility, planning
E
Eni Collaku
Tech & prototype development
A
Anahita Ramesht
User research & testing
03Discovery
Read, then ask
25+ sources split five ways, then one expert interview each. Three of the five told us we were wrong about something.
Social framing beats points and badges. Make it playful, never a competition.
Thiel, 2016 · why the wheel has no score, no loginSK
Silvia-Kay KenderHCI, TU Wien
"Sophisticated formats alienate the people you're trying to reach."
Changed: dropped art-installation ambitions for layered, low-effort engagement.
LJ
Luke JudayUrban planning
"Would you rather" gets priorities. Open questions get wish lists.
Changed: the question format, plus placement: somewhere with footfall already, like a park or café.
AS
Amir H. ShahinSecurity
Temporary IDs give single-use voting without ever knowing who voted.
Changed: killed our planned email-verification step. Not needed.
IH
Ina HysaCommunity office
The real fault line is longtime residents versus new arrivals.
Changed: every vote now ends by showing the neighbourhood's answer.
04Cultural probes
Five painted shoeboxes
Experts explain how participation works in general. Finding out what this specific district feels needed a different method: probe kits, hand-delivered. All five came back.
5
probe kits returned
5
expert interviews
25+
literature sources
3
rounds of testing
💌 Love Letter
Full of belonging and shared identity.
📍 Emotion Map
Clustered on Hannah-Arendt-Platz and the Seepark: our placement.
✅ Preference Checklist
Touchscreen won. Anonymity beat recognition.
🫙 Wish Jar
Became the actual question bank on the wheel.
05Workshop & game
A game as research
A workshop, here, meant the whole team plus our tutor at a glass wall, sticky notes out, working through one question live: how do you turn "give your opinion" into something people actually want to do?
We mapped it as a game: how it starts, how a round plays out, how players decide, how it ends. Then we built the version we liked and tested it on real people.
Dice and debate
Two players argue opposite sides, audience votes a winner. Cut: arguing out loud is uncomfortable for a lot of people.
Board game on a map of Seestadt
Agree on a change, place a token on the map. Cut after tutor feedback: too much setup, too many rules to hold in your head.
Intruder in the District
Three players get the same question, one gets a slightly different one, the group finds the odd answer out.
Playtesting surfaced one problem worth keeping: quieter players struggled to speak up in front of the group. That's the reason the wheel ends in a single private tap, never a spoken answer.
The concept, end of semester one
Video not loading? watch it on YouTube directly.
06Synthesis
Findings into decisions
If a finding didn't change a decision, it was decoration.
Touchscreen beat every input method
→
A plain tablet, nothing else.
Anonymity scored high, competition scored zero
→
No accounts, no leaderboard.
Quiet players got squeezed out of our own game
→
One private tap, never a performance.
Impact motivates, prizes don't
→
"Goes to the city council," not points.
Emotion maps clustered on two squares
→
Placement: Hannah-Arendt-Platz, Seepark.
People want to see their answer land somewhere
→
Every vote ends showing the results.
We also wrote best- and worst-case scenarios around three fictional residents, Alma Meier, Emma Fischer and Jorge Ricardo, walked out in full in the concept video above.
07Testing
Three rounds, before we built it
Sketches, then screens, then the object. Navigation was rarely the problem. Our judgement calls were.
Round 1 · Sketches
First sketch: eight topics
→
Final wheel: four topics
Eight segments looked fine on paper, unreadable at three metres. Four bought bigger type and room for icons.
Round 2 · Wireframe, 7 users
Built
"Backgrounds became the focal point, not the content." — tester
Changed to
Flat colour, doubling as the topic code.
Built
An email step "for vote security." A tester refused to give one.
Changed to
Cut. Walk up, spin, vote.
Round 3 · Cardboard mockup
Icons plus labels, confirmed. Voter counts and "goes to the council," added after testers asked for proof.
I would emphasise once again that my answer will be sent to the city council.
Mockup evaluation participant08Identity
Built from its own parts
C: pie-chart segments, the same shape as the result charts. W: two arcs and the wheel's own pointer.
#1B65A5
Infrastructure
#FBC102
Culture
#F34923
Safety
#29A86B
Greenery
The palette is also the machine interface: the sensor reads these exact colours to know which topic you landed on. Branding decision and hardware constraint, at once.
09The build
What's actually inside it
A colour sensor behind the pointer, a microcontroller, and a screen that reacts. Laser-cut MDF, hand-painted, spun on a screw. My job was feasibility: keeping the concept inside what we could actually build.
Planned
Lazy-susan turntable
Wobbled. Parts list kept growing.
Built
A screw as an axle
Fewer parts, no wobble.
🎡
Spin
Screw axle, dowel handles
👁️
Read
Sensor IDs the colour
🎲
Pick
A random question loads
👆
Answer
Screen recolours, one tap
📊
Compare
Live chart of the votes
RGB colour sensorMicrocontrollerLocal web serverChart.jsLaser-cut MDF
| Component | Role | Source |
|---|---|---|
| Microcontroller | Serves the app, reads the sensor | Lab |
| RGB colour sensor | Reads the segment colour | Ordered |
| MDF sheets | Laser cut wheel + arrows | Lab |
| Monitor stand | Holds the wheel, kills wobble | Second-hand |
| Tablet | Display + input | Lab |
The real thing, spinning
10Final test
No instructions
Five people, told only it was a university project. No explanation of how it worked.
All five worked it out. The shape and the arrows did the explaining. Groups argued before tapping, which is exactly the informal debate the wheel was designed to provoke.
Worked
- Understood without supervision, every time
- Anonymity and one-tap voting praised
- No time limit made people relax and think
Broke
- Page loads lagged a few seconds
- Wheel dragged audibly on the arrows
- One asked if the results were even real
A civic tool lives or dies on whether people believe their input goes somewhere.
Team debrief, after the final test11Outcome
Exhibited in Seestadt
Tablet UI, tech sheet, brochure, A1 poster, all one identity. Pitched live, in German, my third language.
What transfers to digital product
01
Zero onboarding
A stranger, no explanation, one clear action. Same problem as any first-run flow.
02
Hard constraints
Knowing what's buildable, and cutting scope before it becomes a problem.
03
Trust as UX
People give data only when they see where it goes. Same question, any product.
12What I learned
The full process, once
Every stage in order, two semesters, one team. The interesting constraint was never the one in the brief. It was a light sensor and a wooden panel.
🎙️
First expert interviews
Mine was Jan Maly. The follow-up question is where the value hides.
🧰
Resourcefulness
Shoeboxes, reused game cards, cardboard first. Constraint made it better.
🪚
Laser cutting
Designing for a sensor that reads your palette is a different constraint.
🚶
Back to the field
Research stops being abstract once you explain it to a stranger.
🗺️
Two-semester planning
Next time: fewer parallel workstreams, earlier hardware tests.
✂️
Killing good ideas
The turntable, the email step, four of eight logos. Evidence over opinion.
13Archive
Everything else from the process
Material that didn't fit the main story but is still worth a look: an earlier wireframe result screen, a workshop whiteboard or two, the original scenario storyboard, and a couple of research photos.
