Start a conversation
Case study · 15 / 20

Airport Passenger Flow Forecasting (Hybrid AI)

A forecasting system for passenger volume across airport zones, for better queue and staff management.

Airport Passenger Flow Forecasting (Hybrid AI)
Airport Passenger Flow Forecasting (Hybrid AI)
Category
Artificial Intelligence
My role
AI Developer & Model Designer
Date
July 2024
01LSTM

Deep-learning model

02GNN

Graph network

0385%

Forecast accuracy

04Live

Updated forecasts

Challenge

Operations needed to know hours ahead how many passengers each key zone would get, so counters and staff could be arranged and congestion reduced.

Solution

I built a forecasting framework that combines historical passenger data with deep-learning models.

My role

  1. 01

    Data Analysis

    Examined historical passenger-flow patterns.

  2. 02

    Model Design

    Tuned LSTM and graph network models to the problem.

  3. 03

    Implementation

    Implemented and evaluated the forecasting framework.

TechnologiesPythonLSTMGNNPandas

Have a similar project in mind? next project together?