This project is an AI-powered pipeline for predicting and simulating crowd behaviors at large-scale events (concerts, football matches, festivals, parades, etc.). The goal is to help event organizers and safety teams make data-driven decisions for smooth entry/exit, congestion management, and emergency planning.
event-crowd-model/
├── .gitignore # Ignore unnecessary files (e.g., logs, temp, venv, data dumps)
├── data/
│ ├── raw/ # Raw datasets (ticketing, transport schedules, sensors, etc.)
│ ├── clean/ # Processed datasets after cleaning & feature engineering
│ ├── simulations/ # Synthetic/simulated datasets for crowd flow scenarios
├── deployment/
│ ├── app.py # Deployment entrypoint (e.g., FastAPI/Flask app for inference)
│ ├── deploy.sh # Deployment automation script (build, push, run)
│ ├── DockerFile # Docker container definition for reproducibility
├── main.ipynb # Jupyter notebook for exploration, EDA, prototyping & pipeline demo
├── models/
│ ├── congestion/ # Models handling congestion prediction (e.g., bottlenecks, density)
│ ├── forecasting/ # Time-series forecasting models (e.g., arrivals, flow rates)
│ ├── rl/ # Reinforcement learning agents for crowd management strategies
├── README.md # Project overview, setup instructions, usage guide
├── requirements.txt # Python dependencies for environment setup
├── src/
│ ├── model_evaluation/ # Evaluation scripts (metrics, visualization, benchmarking)
│ │ ├── model_evaluation.py # Functions to assess model performance
│ ├── model_training/
│ │ ├── classification.py # Classification models (e.g., crowd state categories)
│ │ ├── reinforcement.py # RL training pipeline (agents, reward design, env interface)
│ │ ├── time_series.py # Time-series models (forecasting arrivals/flows)
│ ├── preprocessing/
│ │ ├── data_cleaning.py # Data cleaning, handling missing values, outlier detection
│ ├── utils/
│ │ ├── helper_functions.py # Utility functions (logging, configs, data loaders, etc.)
1. Clone the repo
git clone "https://github.com/GreatAI-NeuBee/event-crowd-model.git"
cd event-crowd-model2. Create Virtual Environment
python -m venv eventcrowd_env
eventcrowd_env\Scripts\activate
3. Install dependencies
pip install -r requirements.txt
4. Run jupyter notebook
- Open
main.ipynband explore pipeline