flight-look-to-book

This project contains two parts:

1. Data analysis for flight look-to-book statistics, as preprocessing for the machine learning forecast models
2. Forecast models (attempt)

Confluence page: (coming)

Description:

It contains links to the raw datasets, including 'bookings' and 'searches' of the flights worldwide in 2013.

flight-look-to-book-preprocessing_v0.ipynb is the initial version of work, reflecting most of the progress history.

flight-look-to-book-preprocessing_v1.ipynb serves as a cleaner version of solutions to all the 4 questions.

bonus-flight-ltb-forecast-naive-bayesian_v0.ipynb shows the skeleton of forecast, based on the principle of Naive Bayesian. However, it is incomplete yet due to computing power and timing issue.

bonus-flight-ltb-forecast-xgboost_v0.ipynb provides an insight to the Extreme Gradient Boosting Classifier. However, it is incomplete yet due to computing power and timing issue.

Prerequisites:

python --version Python 3.6.8 :: Anaconda, Inc.

pandas (a python library) version: 0.24.2

uname -a Linux 4.15.0-46-generic #49-Ubuntu SMP x86_64 x86_64 x86_64 GNU/Linux

Usage:

flight-look-to-book-preprocessing_v1.ipynb is runnable (yet recommend to SKIP the "Explore" section in solution 4 since it's dependent on hardware performance).