Flight-Delay-Prediction

This repo is 2018 Fall EE608 Final Project: The Prediction of Flight Delays Using Regression Method

Team Member:

Ziran Gong [email protected] Yuqing Luo [email protected] Bowen Li [email protected]

Dataset

Due to the size, I upload all the dataset to google drive. In this project, we use data first month which using the first 3 week to predict the last week.

https://drive.google.com/drive/folders/1LDDwiQW-74P5NFDTEpCAfETxSIBNu7cC?usp=sharing

Web Page

https://sites.google.com/view/ee608

Install

Step1:

1.1 Install Anaconda

1.2 Create Python3.6 environment

conda create -n ee608 python=3.6

Step2:

2.1 Install Jupyter Notebook & JupyterLab

2.2 Install python package

Search and apply the package name below on Anaconda

Or

Using Anaconda Prompt

conda install package-name

• visualization: matplolib, seaborn, basemap
• data manipulation: pandas, numpy
• modeling: scikit-learn, scipy

2.3 Install other package

Search and apply the package name below on Anaconda

• pydot, python-graphviz, pillow

Step3:

3.1 Start JupyterLab

Run code

Download the dataset

Before you run the code, you need to downlown all the 3 datasets using the link and make sure you put the code and datasets in the same folder.

Flight delay prediction

flights-delay-prediction.ipynb

Use Shift + Enter to run code step by step. Then, you can get the result of each process in the middle of the operation. After run it to the final step, you can get the flight delay prediction result for two model.

Airlines Rank & Recommandation

rank.ipynb

Use Shift + Enter to run code step by step. Then, you can get the results in the data processing process. Finally, you can get the rank histogram and the recommandation.

Conclusion

• Data analysis algorithms are applied to predict flight delay.
• Airlines are ranked for recommendation purpose.
• In model 1, cross-validation can avoid bias introduced by splitting data.
• In model 2, compared with linear regression, polynomial regression with ridge regression is the wining method with MSE (54.99).
• Include almost all the factors to rank airline for users.

Notice:

Make sure you download all the datasets and put the same folder before run the code!!!