Airport AI Future Predictor

The Challenge

Airport operations teams at IGIA were making resource decisions reactively — deploying staff and opening counters only after congestion was already visible.

• No predictive system to forecast passenger volume hour by hour

• Manual resource allocation led to overstaffing at off-peak hours and understaffing at peak hours

• Morning rush (6–9 AM) and evening peak (5–8 PM) consistently caused bottlenecks

• No data-driven visibility into check-in counters, security staff, and gate requirements simultaneously

• Holiday and weekend surges were not accounted for in standard planning

The Solution

Groowtth designed and developed a full AI dashboard — combining a trained ML model with a real-time interactive interface.

• Random Forest ML model trained on 4,368 IGIA passenger records — 16 features including hour, day, season, holiday flag, terminal, and weather index

• 95% prediction accuracy with R² score of 0.95 and MAE of just 30.67 passengers

• Interactive 24-hour filter — select any hour and instantly see predicted passengers, counters needed, security staff, and gates to open

• Peak / Off-Peak / All Hours toggle for instant operational view

• Live charts — passenger flow, counters, security staff, and gates across all 24 hours

• Time-based theme — auto-switches between morning, afternoon, and night mode

• Alert system — Peak (400+ passengers), Moderate (250–400), and Low Traffic thresholds with 30-minute advance notifications

• REST API built in — GET predictions per hour, 24-hour forecasts, custom threshold alerts

• Full documentation, model performance analytics, and monthly retraining pipeline