Mood
Viewers explicitly choose how they feel. The product does not infer emotion from private signals.
01Individual UX project
CineGuide
A movie-discovery concept that uses mood, viewing context, and available time to make recommendations easier to choose and understand.
- Role
- UX designer and researcher
- Scope
- Individual course project
- Methods
- Problem framing, interaction design, prototyping
- Deliverable
- Six-screen high-fidelity mobile prototype
The problem
A large catalog does not help when the suggestions ignore the current moment.
Most streaming recommendations rely heavily on viewing history. That can be useful, but it misses immediate needs such as how someone feels, who they are watching with, and how much time they have.
When viewers cannot tell why a title was suggested, they have little information to compare options or correct a poor recommendation. CineGuide reframes discovery as a short, transparent conversation instead of an endless scroll.
The response
Ask for three useful inputs, then show how they shaped the result.
Viewing context
Watching alone, with a partner, or with family changes what a suitable recommendation looks like.
Available time
Runtime becomes a practical filter, so every result can realistically fit the current session.
Recommendation rationale
Mood Match, Context Fit, and Time Fit explain why each movie appears and what the viewer can change.
Prototype evidence
Five artifacts show how the interaction moves from input to informed choice.
Each gallery label explains what the artifact shows and why that decision matters. Use the thumbnails or arrow controls to move through the flow.
Artifact 01
End-to-end prototype flow
Six connected screens move from onboarding through mood, context, time, recommendations, and detailed rationale.
Three explicit inputs keep the viewer in controlRecommendation rationale closes the decision loop
01 / 05
What I delivered
A complete mobile flow with control, feedback, and privacy built into the interaction.
Control: People can change any input, reset preferences, and use lightweight feedback such as “More like this” or “Not today.”
Transparency: A “Why this?” view ties each recommendation to visible contributing factors rather than asking viewers to trust a black box.
Privacy: Mood is treated as session-based data, and longer-term personalization remains optional.