Authors

Ruisi Dai, Yijie Tang, Yuyang Wang

Research Question

How can real-time data be used to predict flight delays at LaGuardia Airport to proactively improve operational decision-making and maximize passenger experience? This project seeks to build a decision-support tool that uses live flight, weather, and alert data to analyze airfield congestion and recommend proactive actions to airport staff. The goal is to prevent excessive taxi times, mitigate delay cascades, and enhance the day-of passenger experience, especially during severe weather events and irregular operations.

Background

LaGuardia Airport (LGA) is one of the nation's busiest and most space-constrained hubs, facing challenges in managing aircraft delays and navigating weather impacts. These bottlenecks not only disrupt operations but directly impact passenger experience, heightening the risk of cancellations, prolonging aircraft taxi times, and congesting terminal areas and roadways.

This capstone project supports the LGA Airport Operations Center (AOC) by delivering a real-time scenario planning tool that predicts flight delays and recommends proactive interventions. The tool helps the AOC minimize downstream disruptions and mitigate negative impacts on passengers and the surrounding Queens neighborhood.

Students are developing a production-ready web application, accessible via desktop and mobile, using Databricks, Python, geospatial analysis, and APIs for flight, weather, and FAA alert data. Working closely with airport staff, the team will translate business rules and operational constraints into real-time logic or predictive models and simulate what-if scenarios for day-of planning.

Methodology

This project builds on the increasing need for real-time operational planning in airside management. The project includes:

• Ingesting real-time and historical data via APIs and internal sources
• Identifying delay and congestion patterns using exploratory and statistical analysis
• Developing logic- or ML-based models to forecast delay risk
• Building a scenario planning interface to test potential interventions
• Deploying a responsive dashboard and app to deliver alerts and insights in real time

Students participate in regular working sessions with airport stakeholders to refine model assumptions, validate results, and design an effective user experience.

Deliverables

• Predictive modeling and analytical pipeline capable of estimating delay risks based on weather and flight activity
• Interactive scenario planner for testing operational decisions
• Responsive web-based dashboard for real-time monitoring of delays and alerts (desktop and mobile-friendly)

Data Sources

• The Port Authority of New York & New Jersey