Tip Predictor Django Application

A complete Django web application that uses machine learning to predict tip amounts based on various dining factors.

Features

• AI-Powered Predictions: Uses XGBoost machine learning algorithm
• Beautiful UI: Modern Bootstrap-based responsive design with animations
• Multiple Factors: Considers 6 key features for accurate predictions:
  • Total bill amount
  • Party size
  • Gender
  • Smoking status
  • Day of the week
  • Time of day (lunch/dinner)

Project Structure

ML_Django/
├── ml_project/              # Machine learning model development
│   ├── 01_ml.ipynb         # Jupyter notebook with model training
│   ├── data/
│   │   └── tips.csv        # Training dataset
│   └── model/
│       └── xgb_model.pkl   # Trained XGBoost model
└── tip_predictor/          # Django web application
    ├── manage.py
    ├── predictor/          # Main app
    │   ├── forms.py        # Django forms with Bootstrap styling
    │   ├── views.py        # Business logic and ML predictions
    │   ├── urls.py         # URL routing
    │   └── templates/      # HTML templates with modern design
    └── tip_predictor/      # Project settings
        ├── settings.py
        └── urls.py

1. Create a Virtual Environment

python -m venv .venv
.venv\Scripts\activate  # On Windows
# or
source .venv/bin/activate  # On Linux/Mac

2. Install Required Libraries

pip install django
pip install pandas numpy scikit-learn matplotlib seaborn
pip install joblib xgboost
pip install ipykernel  # for Jupyter notebook

3. Train the Machine Learning Model

The ML model has already been trained and saved. You can see the complete process in the notebook which includes:

1. Data loading from seaborn tips dataset
2. Feature encoding (categorical to numerical)
3. Model training with XGBoost
4. Model evaluation and comparison
5. Saving the model as xgb_model.pkl

4. Django Project Setup

The Django project has been created with the following structure:

# Project was created with:
django-admin startproject tip_predictor
cd tip_predictor
python manage.py startapp predictor

5. Key Components

Forms (predictor/forms.py)

• Enhanced with Bootstrap styling and form validation
• Includes all 6 required features with proper field types
• User-friendly labels and placeholders

Views (predictor/views.py)

• Handles form processing and ML predictions
• Implements proper feature encoding as per the ML model
• Error handling for model loading and predictions

Templates

• Base template: Modern responsive design with Bootstrap 5
• Home page: Interactive prediction form with real-time validation
• About page: Information about the model and technology stack
• Features: Gradient backgrounds, animations, and modern UI components

URLs

• Clean URL routing with app namespacing
• Integrated with main project URLs

6. Running the Application

1. Navigate to the Django project:

   cd tip_predictor

2. Run migrations:

   python manage.py migrate

3. Start the development server:

   python manage.py runserver

4. Open your browser and visit:

   http://127.0.0.1:8000/
   

7. Using the Application

1. Fill out the prediction form with dining details
2. Click "Predict Tip Amount"
3. View the AI-generated tip prediction
4. Explore the About page for more information

Model Information

• Algorithm: XGBoost Regressor
• Training Data: Restaurant tips dataset from Seaborn
• Accuracy: High performance with low MAE and good R² score
• Features: 6 input features with categorical encoding:
  • Sex: Male=1, Female=0
  • Smoker: Yes=1, No=0
  • Day: Fri=0, Sat=1, Sun=2, Thur=3
  • Time: Dinner=0, Lunch=1

Technology Stack

• Backend: Django 5.2.5
• Frontend: Bootstrap 5.1.3, Font Awesome 6.0
• Machine Learning: XGBoost, Scikit-learn
• Database: SQLite
• Python: 3.8+

Screenshots

The application features:

• Modern gradient backgrounds
• Responsive card-based layout
• Interactive form elements
• Animated prediction results
• Professional navigation
• Information-rich about page

Future Enhancements

• User authentication and prediction history
• Data visualization of predictions
• Model performance metrics dashboard
• Batch prediction capabilities
• API endpoints for external integration