What drives passengers’ choice of airport and airline?

This is my academic project. Data, codes, and report are unable to be published due to privacy.

Problem

What drives passengers’ choice of airport and airline?
There are two international airports in Seoul Metropolitan Area:
  • Gimpo airport: Smaller and old, but closer to the city
  • Incheon airport: Larger and new hub airport, but farther from the city

There are 4 airlines: Korean Air, Asiana Air, Korean LCC, Foreign Carriers
We developed models to investigate passengers' airport and airline choice behavior.

Data description

Data sets: 27 variables from survey data of 488 respondents.
  • Socio-demographic: age, gender, occupation, income, etc.
  • Alternative-Specific : flight information, travel time, mode of transport, etc.

Solution and Results

Developed Descriptive and Predictive Models to Identify Passengers' Behavior Pattern
  • Cleaned data, analyzed descriptive statistics, and visualized Exploratory Data Analysis utilizing Python.
  • Feature engineering: Re-categorizing variables.
    - E.g. Airlines: Korean Carriers vs. Foreign Carriers;
    - E.g. Occupation (re-group 12 categories down to 4 categories): Business, Government, Unemployment and Other).
  • Variables selection: Implemented iterative approach (Backward/Forward selection), conducted Correlation analysis, and used Holdout method to split dataset into training and testing set.
  • Developed logistic regression and decision tree model, yielding the accuracy rate of 84% for airport model and 72% for airline model.