StadiumFlow AI is a GenAI-enabled smart stadium assistant built for the Smart Stadiums & Tournament Experience challenge.
The solution supports fans, volunteers, organizers, and venue staff during large tournaments such as the FIFA World Cup 2026. It helps improve stadium navigation, crowd management, accessibility support, multilingual-style assistance, and real-time operational decision support.
Smart Stadiums & Tournament Experience.
Large stadium events create operational pressure for fans and venue teams. Fans may struggle with navigation, exits, transport, accessibility needs, food zones, queues, and emergency support. Venue staff and volunteers also need fast, clear, and context-aware instructions during crowded situations.
A normal information page is not enough because stadium conditions can change quickly. StadiumFlow AI acts as a decision-support assistant that converts user context and live operational inputs into clear guidance.
The project uses a lightweight decision engine to analyze stadium context.
The system considers:
- User type
- Current location
- Fan need
- Crowd level
- Wait time
- Incident count
- Zone risk
- Accessibility needs
- Volunteer role
Based on these inputs, the system generates:
- Fan guidance
- Crowd-risk labels
- Ranked stadium zones
- Volunteer briefings
- Accessibility-aware movement advice
The app uses transparent rule-based logic to simulate GenAI-style assistance without requiring external APIs or secret keys.
The app contains four assistant modes:
Fans enter their location and need. The assistant generates short, clear guidance for navigation, transport, medical support, food zones, or exits.
Venue staff enter crowd levels, wait times, and incidents for stadium zones. The system calculates a risk score and ranks zones by operational priority.
Organizers can generate short volunteer instructions based on zone, issue type, and risk level.
The assistant creates simple guidance for wheelchair users, elderly visitors, low-vision visitors, families, and lost visitors.
- Crowd level is entered on a 1 to 5 scale.
- Wait time is estimated by venue staff.
- Incident count refers to currently reported local issues.
- The app is a prototype and does not connect to live stadium sensors.
- The system does not collect or store personal data.
The project separates UI code from decision logic. app.py handles the Streamlit interface, while stadium_engine.py contains the scoring and recommendation logic.
The app does not use API keys, passwords, databases, or private user data. It can run locally without secrets.
The app is lightweight and uses simple Python logic. It has minimal dependencies and stays well under the 10 MB repository limit.
The core functions are tested using Pytest in tests/test_stadium_engine.py.
The app includes accessibility-specific guidance for wheelchair access, elderly support, low-vision support, hearing support, families, and lost visitors.
- Python
- Streamlit
- Pandas
- Pytest