CabinIQ

CabinIQ Homepage
CabinIQ Features
CabinIQ Login Page (using Auth0)
Dashboard 2D View
Dashboard 3D View

Inspiration

Air travel can be stressful, and flight attendants often struggle to identify which passengers need immediate attention in a cabin of 200+ people. We were inspired by the challenge of making air travel safer and more comfortable through technology. What if flight crews had a "sixth sense" for passenger wellbeing? We envisioned a system that could transform invisible biometric signals into actionable insights, allowing crews to provide proactive care rather than reactive responses.

What it does

CabinIQ is a real-time passenger stress monitoring platform that combines cutting-edge biometric sensing with AI-powered task management. The system captures heart rate and breathing data from passengers using camera-based photoplethysmography (PPG) technology via Presage mobile app. This data flows to a live dashboard where flight crews can:

Visualize stress levels across the entire cabin with color-coded seat maps (both 2D and interactive 3D views)
Monitor live vitals for individual passengers in real-time
Receive AI-generated task recommendations from Gemini prioritizing which passengers need attention and why
Track stress patterns across different flight phases (takeoff, cruise, landing, turbulence)
Filter passengers by stress level for quick triage

The AI assistant analyzes passenger biometrics and generates detailed, prioritized tasks with explanations of possible causes and recommended crew actions.

How we built it

We designed CabinIQ as a real-time, cross-platform system that bridges mobile biometric sensing with intelligent crew dashboards. The frontend was developed with React, Vite, and Three.js. The backend integrates live passenger vitals from our iOS app using FastAPI REST endpoints, with data validation and deduplication logic to ensure reliability.

The biometric collection layer was built using SwiftUI and the SmartSpectraSwift SDK from Presage, which captures heart rate and breathing data through camera-based photoplethysmography (PPG). We engineered a stress calculation algorithm that combines pulse and breathing metrics with weighted scoring (70% heart rate, 30% breathing rate) to generate real-time stress indices categorized as Calm, Elevated, or High.

On the frontend, CabinIQ integrates with the Gemini 2.5 Flash API, which powers the AI assistant. Gemini analyzes passenger biometric patterns and generates contextualized, prioritized crew tasks with detailed explanations of possible causes and recommended interventions, enabling crews to understand the "why" behind stress levels and act proactively.

Each dashboard module includes a 2D seat map that represents the seating layout of a Boeing 787-9 jet, and you can click each seat to view a detailed set of metrics for that passenger. Crew members can also filter the map by stress level to better help prioritize those who need the most assistance. The 3D model uses Interactive Three.js, turning the cabin map into a movable render to potentially help crew members with finding distressed passengers.

We also created a flight simulation, which highlights the potential stress levels of the passengers at different points of the flight, like takeoff, landing, and turbulence. The simulation fluctuates each passenger's vitals every 3-5 seconds. Auth0 handles secure authentication, ensuring that only authorized crew members can access CabinIQ's sensitive passenger health data and cabin management tools.

Challenges we ran into

Real-time data synchronization: Ensuring seamless flow from mobile sensors to web dashboard required careful polling intervals and deduplication logic to avoid redundant updates
PPG accuracy: Camera-based vital sign detection is sensitive to lighting and movement. We had to implement validation logic to reject impossible readings (pulse/breathing ≤ 0)
3D visualization performance: Rendering an interactive cabin with Three.js while maintaining smooth updates required optimization of geometry and camera controls
AI prompt engineering: Getting Gemini to consistently generate well-structured, actionable tasks required iterative refinement of prompts and output formatting
Multi-platform coordination: Synchronizing iOS, Python backend, and React frontend across network boundaries

Accomplishments that we're proud of

We successfully developed an end-to-end pipeline from biometric capture to AI-driven insights. Integrated non-invasive vital sign monitoring via smartphone camera through Presage. Our intuitive 3D cabin visualization makes complex data immediately understandable. Our stress calculation algorithms accurately categorize passenger states and help make the crew response efficient through valuable Gemini AI insights that generates genuinely useful, context-aware crew tasks.

What we learned

We learned extensively about the operational challenges flight attendants face and what tools would genuinely improve both their workflow and passenger experience. Through research into American Airlines' current crew systems, we discovered that flight attendants value simplicity, real-time information, and actionable intelligence without added cognitive load. This shaped our focus on accurately identifying passenger stress levels and providing clear, prioritized recommendations for different situations. Building an end-to-end healthcare data pipeline that connects real-time biometric collection to a live dashboard taught us the complexities of maintaining data validity, system responsiveness, and operational simplicity during high-pressure flight conditions.

What's next for CabinIQ

Historical analytics: Add trend analysis and stress pattern recognition across flights
Integration with cabin systems: Connect with HVAC, lighting, and entertainment systems for automated comfort adjustments
Wearable integration: Support smartwatches and fitness trackers for continuous, passive monitoring