UX Design · 13 Weeks · Team of 3

Magic Mirror

A smart closet interface that eliminates outfit planning stress through integrated weather, schedule, and wardrobe management.

Final Magic Mirror interface

Photo by JP Valery on Unsplash

Final Design · Mirror Home Screen

The Problem

Planning what to wear is a daily frustration that wastes time and mental energy.

People pull out multiple outfits, change several times, and still walk outside unprepared for the weather or day's activities.

Design Goal

Create a smart mirror interface that displays closet, weather, and schedule in one glanceable view.

Project Details

Role

UX Designer

Duration

2 Weeks

Team

3 Designers

Tools

Figma, ChatGPT

Key
Considerations

01

Weather

What's the temperature and forecast? Will I need an umbrella or jacket?

02

Schedule

Work meeting, job interview, casual dinner, or gym session?

03

Preferences

Personal style, comfort levels, and what makes you feel confident.

04

Availability

What clothes are clean and accessible in your closet right now?

Key Insight

While weather apps and closet apps exist separately, no solution combines wardrobe, weather, and calendar in a single glanceable interface.

Competitive
Analysis

We analyzed existing solutions across four categories: smart mirrors (Lululemon Studio Mirror), closet management apps (Indyx), outfit styling services (Stitch Fix), and smart closets (LG Styler).

None of them combined wardrobe, weather, and calendar in a single glanceable interface. That gap became our design goal.

Smart Mirrors

Lululemon

Fitness focus, no wardrobe integration

Closet Apps

Indyx

Wardrobe only, no weather or calendar

Styling Services

Stitch Fix

Outfit help, but no real-time context

Smart Closets

LG Styler

Cleans clothes remotely, no outfit planning

Initial
Exploration

Sketching allowed us to explore concepts rapidly without worrying about pixel perfection. The focus was on ideas, not polish.

Whiteboard sketches exploring phone vs mirror

Phone vs. mirror — early concept exploration

Table sketches exploring screen structure and navigation

Screen structure and navigation ideation

Key Questions

  • • Mirror or mobile app?
  • • Individual items vs. complete outfits?
  • • Essential vs. nice-to-have features?
Design Challenge

The Flatlay Problem

My initial approach was a "flatlay" layout—displaying clothes in the order they're worn (hats → jackets → pants → shoes). It felt immersive and natural.

The problem: I couldn't make it work aesthetically. The spacing felt awkward, and it didn't scale across different screen sizes.

Original flatlay attempt

Original approach - flatlay layout

Breakthrough

Team Collaboration Wins

My teammate Shelby had designed the mobile app using a horizontal scrolling card layout. After reviewing her work, I realized this approach was:

  • More flexible for different screen sizes
  • Easier to scan quickly
  • Consistent across platforms

"Even though the flatlay was more immersive, cross-platform consistency was more valuable."

Shelby's mobile design

Shelby's card-based approach

Usability
Testing

Testing with 5 users revealed critical usability issues we never anticipated.

01

Invisible '+' Button

0/5 users could find the button to add clothing without prompting—it was buried at the end of a scrollable row.

02

Hidden Scrolling

Users didn't realize the clothing row scrolled horizontally—they thought what they saw was everything.

03

Confusing Calendar

Calendar icons and date indicators were misinterpreted—users couldn't distinguish their meaning.

04

Messy Weather Layout

Users confused hourly and daily forecasts, treating the two-column layout as a single dataset.

"I love getting feedback like this because it's like a light comes on when I'm designing. I can see! I know what the users need."

Iterating on
Feedback

Solution 1

Making the '+' Button Discoverable

Before
Before - hidden button

Button buried at end of scrollable row

After
After - prominent button

Prominent '+' button at the front (inspired by Facebook Stories)

Impact: 5/5 users successfully added clothing items in follow-up testing.

Solution 2

Redesigning the Weather Interface with AI Assistance

ChatGPT's Recommendations

  • • Add a prominent current conditions section at the top
  • • Make hourly and 10-day forecasts horizontally scrollable
  • • Use distinct visual styling to differentiate forecast types
Original weather

Original: Confusing two-column layout

ChatGPT suggestions

ChatGPT's suggestion mock-up

Final weather

Final: Clear hierarchy and scrolling

Impact: Weather comprehension time dropped from 45+ seconds to under 10 seconds.

Final
Design

Final mirror interface

Prominent '+' Button

Easy clothing addition

Scrollable Preview

Browse full outfit

Weather At-a-Glance

Current conditions prominent

Quick Access

Settings & detailed weather

Impact

Before Iteration

0/5 users found the '+' button
45s to understand weather layout

After Iteration

5/5 users successfully added clothing
10s to scan weather information

"I would actually use this every morning—it solves a real problem I have."

— Usability Test Participant

"The weather layout makes so much more sense now. I can see what I need immediately."

— Usability Test Participant

Key
Takeaways

01

AI as Design Advisor

ChatGPT helped me identify established UI patterns I hadn't considered, but I still needed to validate suggestions, adapt to our context, and make final decisions based on usability principles.

Learning: AI tools accelerate research and ideation, but human judgment is essential for contextual design decisions.

02

Collaboration Over Individual Vision

I was attached to the flatlay concept because it felt "innovative," but Shelby's card layout was clearly superior when I evaluated both objectively. Setting ego aside led to a better solution.

Learning: The best design isn't always the most novel—sometimes it's the one that works seamlessly across platforms.

03

Testing Reveals the Invisible

The '+' button was perfectly visible to me because I designed it. The weather columns made sense to me because I created them. Users saw neither.

Learning: Designer's curse of knowledge is real. Even informal testing reveals critical blind spots.

What I'd Do
Differently

Conduct Upfront Research

Interview 5-10 users about outfit planning to reveal needs we missed when relying on our own assumptions.

More Low-Fi Iteration

Gray-box wireframes would have let us test more layout variations before committing to high-fidelity.

Test Physical Context

Test with users standing in front of an actual mirror to reveal distance, angle, and lighting considerations.

Explore Input Methods

Design the wardrobe management flow—how users photograph and categorize their clothes.