✈️ Airlines Delay Flight Prediction using Machine Learning

Project Overview

This project predicts whether a flight will be delayed or on time using various Machine Learning algorithms.
We explore flight delay data, perform data preprocessing, train multiple classification models, and evaluate them using accuracy, ROC curves, and feature importance.

---

Using Requirements

pandas, numpy, matplotlib, seaborn, scikit-learn, xgboost

---

Dataset

• Source: Airline on-time performance dataset (Kaggle-https://www.kaggle.com/datasets/ulrikthygepedersen/airlines-delay)

• Features:

  • Airline – Airline carrier code
  • Flight – Flight number
  • Origin – Departure airport
  • Destination – Arrival airport
  • Distance – Distance between airports
  • Scheduled_Departure – Scheduled departure time

• Target:

  • Delayed – 1 if delayed, 0 if on-time

---

Workflow

1. Import Libraries
2. Exploratory Data Analysis (EDA)
   • Summary statistics
   • Class distribution check
   • Outlier detection
   • Correlation heatmaps
3. Data Preprocessing
   • Handling missing values
   • Encoding categorical features
   • Normalization/Scaling
4. Model Training
   • Decision Tree Classifier
   • Random Forest Classifier
   • Logistic Regression
   • XGBoost Classifier
   • AdaBoost Classifier
5. Model Evaluation
   • Accuracy Score
   • ROC Curve Visualization
   • Feature Importance Plot

---

Results

Model	Accuracy
Decision Tree	68%
Random Forest	70%
Logistic Regression	58%
XGBoost	66%
AdaBoost	62%

---

Key Visualizations

• Class Distribution

• Correlation Heatmap

• ROC Curves

• Feature Importance (Random Forest)

Run Jupyter Notebook

jupyter notebook Airlines_Delay_Flight_Prediction_using_ML.ipynb