🔗 Transmedia Project — Sort It Out is the interactive AR node in a larger campaign turning recycling from a chore into a narrative.
▸ Case Study 02

Sort It
Out

Mobile Web Game AR Overlay Behaviour Change Transmedia · 2026

A mobile recycling game triggered by scanning a QR code at a campus bin — intercepting the student at the exact moment the disposal decision happens.

2 Prototypes
590 Positive Training Images
80 Training Epochs
75% Confidence Threshold
20% Target Improvement
5s Decision Window

Step 01

The disposal decision: primary and secondary research

This project builds on primary research from EMDD 620 (Ava, Femi, and Selasi), who identified contamination patterns at Ball State waste stations. The EMDD 640 team extended that work through observation, interviews, and journey mapping.

Ball State recycles paper, cardboard, glass, metals, and specific plastics in blue bags. Despite this infrastructure, research found one consistent pattern: students slowed at bins but rarely read signage. Sorting was happening in a near-total information vacuum.

Method What we found
Observation (Woodworth, DeHority, Student Center) Students paused at bins but rarely read signage. Greasy pizza boxes were the single most common contamination source.
Interviews with Sustainability Officers + Republic Services Contamination rates high enough to divert entire loads to landfill. Muncie facility enforces a strict 10% threshold.
Journey mapping (~10 students) Confusion peaks in rushed between-class moments. The disposal decision happens in under five seconds.
Discourse analysis (Ball Bearings, Reddit) Students frustrated by unclear rules, inconsistent bin placement, and scepticism that recycling is actually processed.
Secondary research (Noh, 2021; Cho, 2019; Yusoff et al., 2024) Recycling participation driven by perceived behavioural control and clear guidance, not general awareness. The intention-action gap is widest at the point of disposal.
Students broadly knew recycling mattered. They did not know, at the moment of holding something over a bin, what Ball State's specific rules were — and existing infrastructure wasn't telling them.

Step 02

Not an awareness problem. A narrative coherence problem.

Existing tools operate at the wrong moment. Posters and orientation modules build awareness weeks before the disposal decision. They cannot intervene at the bin. Sort It Out is designed to.

▸ How Might We How might we design a transmedia experience that uses narrative, interaction, and participation to reshape how students understand recycling — meeting them at the exact moment the disposal decision happens?
Site-Specific
The game at Woodworth is not the game at DeHority. Each location carries narrative text developed with its community.
Mobile-First, No Download
Students are on their phones at the bin. A QR code triggers a mobile browser experience instantly.
Consequence-Based
Wrong sorts trigger an explanation before the item resets. Every error is a teaching moment, not a failure state.
Part of a Storyworld
The spatial touchpoint connecting the transmedia campaign's digital and real-world layers.

Step 03

Sort It Out as one node in a larger storyworld

The project is structured as a transmedia campaign. Sort It Out is the spatial node — the only component that can intercept a student at the physical bin.

01
Documentary
How Does BSU Recycle? — the narrative anchor
02
Explainer Series
TikTok and Reels — spreadable awareness content
03
Sort It Out
AR game at the bin — this project
This Project
04
Street Quiz Videos
Gamified, shareable campus content
05
Recycling Leaderboard
Residence hall competition app — sustain layer
01
Scan QR code on the bin — location-specific splash screen appears
02
Sort 8 waste items — wrong sorts trigger a one-line explanation
03
Stamp the location — floor climbs the campus leaderboard
04
Complete all locations — earn a dining coupon reward

Step 04

The mechanic, demonstrated

Two prototypes, each extending the last. Prototype 1 proved the mechanic. Prototype 2 proved the mechanic could know where it was.

Prototype 1 · Core Mechanic
Slingshot sort interaction — Angry Birds-style
Mobile-first, in-browser, no download
Wrong sorts trigger explanation before reset
Confirmed mechanic was legible in a between-class moment
Prototype 2 · Live AR Build
Custom-trained CV model detects real bin via camera
Replaces placeholder images with player's actual scanned bin
Compass-bearing spatial overlay — items hidden at randomised angles
Perspective illusion via phone tilt — body does the sorting
Documentary video reward closes the loop
▸ Prototype 1 · Core Mechanic Web build · designed for mobile

This game is designed for mobile. Open it full screen for the best experience.

Open full screen ↗
▸ Prototype 2 · Live AR Build Live camera · custom CV model · real bin detection

This prototype uses your phone camera to detect a real bin. Open it full screen for the intended experience.

Open full screen ↗
590 Positive bin images
1,049 Negative images
80 Training epochs
75% Confidence threshold

Runs entirely on-device via TensorFlow.js. Outside prototype scope: geofencing, QR triggering, live leaderboard backend, location-specific item sets.


Step 05

Testing plan

Formative Testing · Spring 2026
Think-aloud walkthroughs, 5–8 students per round
Concept testing: Does this feel relevant to your daily life?
Early social media analytics from Explainer Series — saves and DM shares as leading indicators
Summative Testing · Spring 2027
Pre/post surveys: recycling confidence, correct item ID, self-reported habits
Physical contamination data from Facilities + Republic Services vs Fall 2026 baseline
App analytics: sorting completions, contamination score changes, leaderboard performance
Exit interviews with 10–15 students on story elements and behaviour change
The test is not whether players found the game fun. The test is whether they know what to do the next time they stand at a bin holding a greasy cup.

Step 06

What changed, and what still needs to

The hardest design problem in Sort It Out is not the recycling taxonomy. It is finding the balance between flow state, convenience, and education — three things that pull against each other. Push too hard on education and the game feels like a quiz at the worst moment. Optimise entirely for flow and students complete a round without retaining anything useful.

The first prototype solved the mechanism. The video reward — a documentary clip showing where a sorted item actually goes — was the response to professor feedback naming the gap: the game needed to mean something beyond itself. The sort earns the story; the story earns the next sort.

The first prototype proved the mechanic works. The next one has to prove it changes something.