×
Community Blog Project Showcase | Global Flight Delay Prediction Using Machine Learning

Project Showcase | Global Flight Delay Prediction Using Machine Learning

This project is from the team MnoL, which was awarded with the Third Prize in the Global AI Innovation Challenge 2021 - Intelligent Weather Forecast for Better life.

This project is from the team MnoL, which was awarded with the Third Prize in the Global AI Innovation Challenge 2021 - Intelligent Weather Forecast for Better life.

Project Introduction

This project uses regression techniques of machine learning to predict the flight delay. By predicting the flight delay beforehand, the airport can reduce their cost. We took several independent variables such as humidity, pressure and precipitation for predicting the final delay.

1

Technology Highlights

In this project we developed a high level customizable API which could be used to predict flight delay on any airport around the world. Initially, the model trained itself by scraping past weather data and flight details of the given airport. Then, it containerized itself and set up a production grade flask server incorporated with nginx and gunicorn which is further deployed to Alibaba's cloud service and published in Alibaba's API marketplace.

Moreover, the process mentioned above is automated with the use of Alibaba workflow in this github repository, so that during the official release the changes could easily be reflected in the production API without any delay or overhead.

During construction of this project, the main focus was on flexibility, so that it could easily be used by professionals and beginners alike. Moreover, we have tried to keep the API as user-friendly as possible while maintaining its robustness.

The developed API can later be used for a multitude of purposes, such as:

  • Creating a mobile/web app which depicts flight weather delays to customers with a very high accuracy
  • Integration with airline maintenance systems to predict optimum time for flight maintenance
  • Integration with Air Traffic Control (ATC) for better management
  • Integration with airline booking system to increase efficiency

Alibaba Cloud Products Used

Container Registry was used for containerizing and hosting our final docker image.

ACK was used in our repository's workflow for continuous integration and deployment

ECS was used for hosting and deploying our API on independent apache server.

API Gateway was used to publish our API in Alibaba's API marketplace.

About the Developer

We are students, currently in the penultimate year of our engineering. We have been actively working in the field of ML and AI. We have also explored various subdomains such as NLP, time series analysis etc.

0 0 0
Share on

Alibaba Cloud Project Hub

94 posts | 15 followers

You may also like

Comments

Alibaba Cloud Project Hub

94 posts | 15 followers

Related Products