Mood Mate
A minimal, camera-based mood detection app that pairs facial-expression analysis with a moment of real self-reflection
Early demo — UI and detection flow still evolving
Project Overview
Most mood trackers rely on people remembering to log how they feel — which is exactly the habit most people struggle to keep. Mood Mate takes a different approach: it reads facial expression through the camera in the moment, then turns that reading into a short act of self-reflection instead of a one-line confirmation.
The goal isn't just detection — it's emotional self-awareness. The app pairs a quick, low-friction mood check with a "how are you really doing" prompt, so the interaction feels closer to a moment of pause than a data-entry task.
Key Features
- Personalized home screen — Time-aware greeting ("Good morning") with a "How are you today?" prompt
- Camera-based mood check — Live facial expression analysis through a custom-shaped capture UI
- Mood result card — Bottom sheet showing detected mood (Happy / Neutral / Tired, etc.) with a smile-confidence percentage
- Reflection step — "My mind right now" prompt after saving, in place of a plain confirmation toast
- Mood-aware suggestions — Activity recommendations tailored to the detected emotional state
Designing the Save Flow
A small change with an outsized effect on how the app feels to use.
Generic confirmation
"Mood Saved! Your mood has been successfully recorded." — closes the loop, but asks nothing of the user.
"My mind right now"
A short, optional reflection prompt before "Done" — turning a save action into a moment of awareness.
Design note: the original toast technically worked, but it treated mood-logging as a data event. Replacing it with a short reflective prompt keeps the same number of taps while making the moment mean something to the user.
Exploration: "Mind Reader"
Concept / R&DAlongside the camera-based flow, I prototyped a more experimental direction for Mood Mate: a question-driven "Mind Reader" mode that infers mood through a short flow of questions rather than the camera alone.
It explores a Bangladesh-specific name database for lightweight personalization, a candidate-scoring
system to weigh possible mood outcomes, and a set of dedicated Kotlin ViewModel
and scoring-algorithm classes to keep the logic testable and separate from the UI.
Still an early exploration rather than a shipped feature — kept here because it shaped how the main detection flow and its scoring logic were eventually structured.
Roadmap
Journal / Mood History
PlannedA log of past mood checks and reflections
Analytics
PlannedMood patterns and trends over time
Settings
PlannedPreferences, reminders, and privacy controls
Brand identity
PlannedA dedicated Mood Mate logo and minimal visual system
Tech Stack
Project Info
Explore More Projects
Check out my other work