My Fellowship.AI Challenge

Airline On-Time Arrivals - Binary classification model

Pierre Megret

The Goal

To use the US Dept. of Transportation on-time arrival data for non-stop domestic flights by major air carriers to predict arrival delays with a binary classification model.

The data set

US Dept. of Transportation on-time arrival data for non-stop domestic flights by major air carriers. The data set corresponds of the month of January 2016, so that I could test my model on all the year. To predict whether a flight will be delayed or not, here's the features I chose from the USDoT: Day of Week, Unique Carrier, Flight Number, Origin and Destination airport Id and cities Id, CRS departure and arrival times, ARR_DEL15 dummy, DIVERTED dummy, Air time, and Distance.

The Code

The code is organized as follow: an exploration of the data, a grid search cross-validation error to determine the best hyper-parameters of my model, and the predictions. All in Jupyter Python Notebook.

Part 1. Exploration of the data

Looked at the features of the data as well as the existence of missing values, did some data cleaning.

Part 2. Building of the Model

Built a Random Forest Classifier after having tuned it with a grid search cross-validation.

Part 3. Predictions and model performances

Evaluated the model performances on different months, on different years. (February, and October 2016, January 2017)