🛫 FlightDelayML – Flight Delay Detection using Machine Learning

⏱️ A machine learning project that predicts flight arrival delays and classifies flights as delayed or on-time based on various delay factors.

---

🚀 Features

• 📉 Regression Modeling using Linear Regression to predict numeric arrival delay
• ✅ Classification Modeling with Decision Tree to detect delayed flights (accuracy: 98%)
• 📊 Visualization Dashboard with scatter plots, feature importance, and confusion matrix
• 🧼 Data cleaning, encoding, and feature engineering for improved model performance
• 🧪 Evaluated using R², MAE, MSE for regression and accuracy/F1-score for classification

---

📌 Technologies Used

Component	Tool/Library
Language	Python 3.10
ML Models	LinearRegression, DecisionTreeClassifier
Data Handling	pandas, NumPy
Visualization	Matplotlib, Seaborn
Evaluation Metrics	scikit-learn (MAE, R², accuracy, F1-score)

---

⚙️ Installation

git clone https://github.com/akasha456/Flight-Delay-Detection
cd Flight-Delay-Detection
pip install -r requirements.txt

---

🧠 How It Works

flowchart TD
    A[Load Flight Dataset] --> B[Preprocess & Clean Data]
    B --> C[Train Regression Model]
    B --> D[Train Classification Model]
    C --> E[Predict Arrival Delays]
    D --> F[Classify Flights as Delayed/On-Time]
    E --> G[Evaluate Regression Metrics]
    F --> H[Evaluate Accuracy and Confusion Matrix]
    G --> I[Visualize Predictions]
    H --> I

Loading

---

📊 Model Evaluation Snapshot

🔷 Regression (Arrival Delay Prediction)

Metric	Score
R² Score	0.972
MAE	6.97
MSE	90.12
Explained Variance	0.972

🔶 Classification (Delayed or Not)

Metric	Score
Accuracy	98%
Precision	1.00 (Not Delayed), 0.95 (Delayed)
Recall	0.97 (Not Delayed), 1.00 (Delayed)
F1-Score	0.98

---

📈 Feature Importance (Decision Tree Classifier)

Feature	Importance
NAS_Delay	0.5882
Dep_Delay	0.4118
Others	0.0000

---

🌐 Future Enhancements

• ✈️ Integrate live flight data via airline APIs
• 📍 Add geographical visualization of delays by airport
• 🧠 Explore ensemble models (Random Forest, XGBoost)
• 🗂️ Summarize delays by day, airline, or region
• 📱 Build a simple UI for user input and results visualization

---

📜 License

This project is licensed under the MIT License.

---

💬 Acknowledgements

• Scikit-learn for ML algorithms
• Matplotlib and Seaborn for visualizations
• Kaggle for access to flight datasets

---