maco-data/airline-fare-prediction

EDA and ML practice case

Airline Price Predictor

EDA & ML Practice

Checklist

Cleaning and Processing of Data

• Import libraries

• Load dataset

• Check for missing values

• Drop unnecessary features

• Rename features

• Reorder featuress

• Change data types

  • Change flight_date to "datetime" in yyyy-mm-dd format
  • Change flight_dep and flight_arr to "datetime" in HH:MM format
  • Change total_stops to "int" from categorical values
  • Change flight_time to "datetime" in HH:MM format
  • Change flight_fare to "int"

• Fix reapeated values in some features

  • Fix feature airline_name, i.e,: "Air Asia" to "AirAsia"
  • Fix features flight_dep and flight_arr, i.e,: "BOM" to "Mumbai"

• Encode features for data Analysis

  • Extract months and days from flight_date
  • Extract hours and mins from dep_time and arr_time
  • Extract hours and mins from flight_time
  • Convert airline_nameto numerical data
  • Connvert flight_dep to numerical data
  • Convert flight_arrto numerical data

Exploratory Data Analysis

• Considering simplifying features name i.e.: flight_fare to price, etc.

  • Change flight_fare to price
  • Change flight_time to duration
  • Change airline_name to airline
  • Change flight_dep to dep
  • Change flight_arr to arr

• Create test data

  • Split data
  • Feauture selection for prediction

• Hypertune Model

  • Using Random Search Cross-Validation
  • Using Grid Search Cross-Validaton

Price Prediction using Machine Learning