The problem AirAware solves

Air travel, while statistically safe, can suffer from rare but catastrophic incidents—like MH370—that stem from anomalies such as technical failures, bad weather, or human error. Traditional systems often *fail to detect these anomalies in real-time, leading to *delays, accidents, or disappearances.

Our project uses Machine Learning-based Anomaly Detection to:

• Identify unusual flight patterns or behaviors in real time.
• Predict potential issues before they occur, using historical data.
• Alert pilots and control centers early, improving communication and decision-making.

This ensures safer, more efficient flights while restoring public trust in air travel.

Challenges we ran into

• Data Preprocessing: Cleaning and formatting aviation data (like altitude, speed, coordinates) was time-consuming.
• *Model Accuracy: Initially, our anomaly detection model gave too many false positives. We solved this by tuning hyperparameters and using *Isolation Forest + PCA together.
• *Visualization Lag: Rendering large volumes of flight data caused performance issues in the browser. We optimized it by switching to *LeafletJS and limiting the live feed size for smoother UI/UX.