Predict circuit risk before incidents occur.
Real-time AI-powered safety intelligence for motorsport officials.
Yellow Flag is a demonstration project created for the IBM AI Builders Hackathon.
The platform showcases how AI can assist motorsport safety operations using weather intelligence, historical incident analysis, and predictive risk assessment.
Data, predictions, recommendations, and risk scores shown within the demo may be incomplete, simulated, inaccurate, or unsuitable for real-world race control decisions.
This project should be considered a proof of concept and not an operational safety system.
Yellow Flag is an AI-powered motorsport safety platform designed to help race directors, stewards, and circuit officials identify elevated risk conditions before incidents occur.
Instead of reacting to crashes after they happen, Yellow Flag continuously combines:
- 🌦️ Live weather telemetry
- 📄 Historical incident reports
- 🏎️ Circuit-specific risk patterns
- 🤖 AI reasoning and recommendations
to generate real-time safety intelligence across an entire race circuit.
Every corner receives its own continuously updated risk score, allowing officials to understand exactly where, why, and when danger is increasing.
Visualise the entire circuit with real-time risk scoring.
- Green / Yellow / Red risk indicators
- Corner-by-corner analysis
- Historical incident concentration
- Weather contribution breakdown
- AI-generated safety assessment
Officials can instantly identify hotspots requiring attention.
Understand how circuit risk evolves throughout a race weekend.
Track projected risk across:
- Practice Sessions
- Qualifying
- Race Day
This allows proactive planning and strategic deployment of safety resources before conditions deteriorate.
Live weather telemetry feeds directly into the risk engine.
Factors include:
- Rain Probability
- Wind Speed
- Visibility
- Temperature
- Weather Trends
Weather conditions are evaluated at a corner level rather than as a circuit-wide average.
Historical motorsport incident reports are transformed into structured safety data using IBM Docling.
Yellow Flag identifies:
- High-risk corners
- Frequent incident zones
- Condition-specific accident patterns
- Historical severity trends
This provides context beyond what current weather alone can explain.
Powered by IBM Granite.
The AI reasons across:
- Current weather
- Historical incidents
- Circuit characteristics
- Existing risk scores
to generate:
- Actionable recommendations
- Confidence scores
- Prioritised alerts
- Safety justifications
The goal is explainable AI rather than black-box predictions.
Prepare for changing conditions before they happen.
Officials can adjust:
- Rain probability
- Wind speed
- Visibility
- Temperature
and immediately see how risk changes across every corner.
Perfect for scenario planning and pre-session safety briefings.
Weather API
│
▼
Risk Engine ───────────────┐
│
Historical Reports │
(IBM Docling) │
│ ▼
└──────────► IBM Granite
│
▼
AI Recommendations
│
▼
Real-Time Safety Dashboard
- Python
- FastAPI
- Next.js
- React
- Tailwind CSS
- IBM Granite (via Ollama)
- IBM Docling
- WeatherAPI
Monitor evolving circuit risk and receive AI-backed safety guidance.
Understand incident-prone locations and prepare interventions proactively.
Allocate marshals, recovery vehicles, and resources more effectively.
Improve situational awareness during changing weather conditions.
Motorsport safety systems traditionally react to incidents.
Yellow Flag aims to shift the paradigm toward:
Predict → Prepare → Prevent
By combining operational data, historical intelligence, and explainable AI, race officials gain a clearer understanding of emerging risks before they become incidents.