LUXORA AI

LUXORA AI is a web-based immersive luxury retail platform designed to explore how augmented reality and behavioural analytics can enhance digital luxury shopping. It combines interactive 3D product visualization, voice-controlled navigation, product storytelling, behavioural analytics, and ML to create an immersive luxury retail experience.

Live Project

Launch LUXORA AI ··· View Analytics Dashboard

Users can:

• Explore luxury products through interactive 3D visualization
• Navigate products using voice commands
• Explore product journeys and provenance stories
• Save products and interact with their selections

System Architecture

                    ┌─────────────────────┐
                    │     LUXORA AI       │
                    │   Web Application   │
                    └──────────┬──────────┘
                               │
              ┌────────────────┼────────────────┐
              │                │                │
              ▼                ▼                ▼
        3D Product       Voice Interface    User Events
        Visualization    Web Speech API     & Feedback
              │                │                │
              └────────────────┼────────────────┘
                               ▼
                    ┌─────────────────────┐
                    │      Supabase       │
                    │   PostgreSQL Data   │
                    └──────────┬──────────┘
                               │
                               ▼
                    ┌─────────────────────┐
                    │  Analytics Dashboard│
                    │       Vercel        │
                    └──────────┬──────────┘
                               │
                               ▼
                    ┌─────────────────────┐
                    │    ML Analysis      │
                    │    & Prediction     │
                    └─────────────────────┘

Technology Stack

Frontend	React · Vite · Tailwind CSS · Framer Motion
3D / AR	Three.js · React Three Fiber · React Three Drei · GLB · Draco
Voice Interface	Web Speech API
Data & Backend	Supabase · PostgreSQL
Analytics & ML	Python · pandas · scikit-learn · TextBlob
Deployment	Cloudflare Pages · Vercel

Behavioural Analytics

The system tracks interactions such as: Product Rotation · Zoom · Product Switching · Journey Views · Category Filtering · Voice Commands · Add to Cart · Ratings · Checkout Completion

Machine Learning

The project explores whether behavioural interactions within an immersive retail environment can be used to predict cart conversion. Models evaluated include:

Logistic Regression ··· Random Forest

The Random Forest model achieved a reported 81.3% accuracy and 0.746 AUC on the project's evaluation dataset.

Project Results

Metric	Result
User sessions	79
Interaction events	4,437
Feedback submissions	24
Average rating	4.6 / 5
Cart conversion	29.1%
Random Forest accuracy	81.3%
Random Forest AUC	0.746

---

Authors: Narda Zaria and Valiveti Aiswarya

⋆｡°✩｡°⋆

_{Note: These results are based on the project's prototype evaluation dataset collected from real interactions of prototype users.}