📱 SmartPredict

Predictive Modeling for Smartphone Purchase Behavior

An interactive machine learning dashboard that predicts smartphone purchasing behavior
based on demographic factors, user preferences, and market trends.

---

---

✨ Highlights

🎯 What It Does Predicts smartphone purchase probability using ML Analyzes demographic patterns & market trends Compares brands with comprehensive metrics Visualizes data with interactive Chart.js charts

🚀 Why It's Great 87%+ accuracy Random Forest model Real-time predictions with instant feedback Beautiful UI with dark mode & animations Production-ready deployed on Azure

---

🎯 Features

📊 Dashboard Sections

Section	Description
Overview	Project metrics, model performance summary, and quick stats
Demographics	Interactive demographic analysis with purchase patterns by age, gender, income
Prediction Tool	Real-time ML predictions with user demographic input
Brand Comparison	Comprehensive brand analysis, market share, and loyalty metrics
Insights	Advanced analytics, user segments, and business intelligence
Documents	Access reports, notebooks, and project documentation

⚡ Key Capabilities

• ✅ 87%+ Prediction Accuracy with optimized Random Forest
• ✅ Interactive Visualizations using Chart.js
• ✅ Real-time Predictions with instant probability scores
• ✅ Brand Intelligence with market share analysis
• ✅ Export Functionality for charts and data
• ✅ Mobile Responsive design for all devices
• ✅ Dark Mode with smooth animations
• ✅ PWA Support with offline capabilities

---

🛠️ Tech Stack

Python

Flask

Scikit-learn

HTML5

CSS3

JavaScript

Azure

Layer	Technologies
Backend	Python 3.11, Flask 2.2.5, Flask-CORS, Gunicorn
ML/Data	Scikit-learn 1.7.1, Pandas 2.3.2, NumPy 2.3.3, Joblib
Frontend	HTML5, CSS3, JavaScript ES6+, Bootstrap 5
Visualization	Chart.js 4.x, Custom animations
Deployment	Azure App Service, Gunicorn WSGI

---

🚀 Quick Start

Option 1: Visit Live Demo

👉 https://smartpredict-app-a3gcecfectcudbdd.centralindia-01.azurewebsites.net

Option 2: Run Locally

# Clone the repository
git clone https://github.com/Mandar123454/Predictive-Modeling-for-Smartphone-Purchase-Behavior-ML.git
cd Predictive-Modeling-for-Smartphone-Purchase-Behavior-ML

# Create virtual environment
python -m venv .venv

# Activate virtual environment
# Windows:
.venv\Scripts\activate
# Linux/Mac:
source .venv/bin/activate

# Install dependencies
pip install -r requirements.txt

# Run the dashboard
python run_dashboard.py

Then open http://localhost:5000 in your browser.

Option 3: Windows Quick Setup

# Just run the setup script
.\setup.bat

---

📁 Project Structure

📦 SmartPredict
├── 📄 app.py                    # Root entry point
├── 📄 run_dashboard.py          # Cross-platform launcher
├── 📄 requirements.txt          # Python dependencies
├── 📄 Procfile                  # Azure/Heroku deployment
│
├── 📂 Dashboard/                # Web application
│   ├── 📄 app.py               # Flask backend & API
│   ├── 📄 index.html           # Main dashboard UI
│   ├── 📂 css/                 # Stylesheets (dark mode, animations)
│   ├── 📂 js/                  # JavaScript modules & charts
│   └── 📂 data/                # Static visualization data
│
├── 📂 Data/                     # ML datasets
│   ├── 📄 smartphone_purchased_data.csv
│   ├── 📄 smartphone_purchased_data_cleaned.csv
│   └── 📄 smartphone_purchased_data_updated.csv
│
├── 📂 Models/                   # Trained ML models
│   ├── 📄 model.pkl            # Random Forest model
│   ├── 📄 scaler.pkl           # Feature scaler
│   └── 📄 model_columns.pkl    # Column mappings
│
├── 📂 Notebook/                 # Jupyter notebooks
│   ├── 📓 Main Notebook.ipynb
│   ├── 📓 exploratory_analysis.ipynb
│   └── 📓 feature_influence_analysis.ipynb
│
└── 📂 Project Report/           # Documentation
    ├── 📄 DOCS.md
    └── 📄 PROJECT_OVERVIEW_ALL_IN_ONE.md

---

🚀 Deploy to Azure

This project is Azure App Service ready with included Procfile and Gunicorn configuration.

# Quick deploy with Azure CLI
az webapp up --name your-app-name --resource-group your-rg --runtime "PYTHON:3.11"

📖 See AZURE_DEPLOY.md for detailed deployment guide.

---

📊 Model Performance

📈 Metrics Metric Score Accuracy 87.3% Precision 85.1% Recall 89.7% F1-Score 87.3% ROC-AUC 0.924

🔬 ML Pipeline Data Preprocessing - Missing values, outliers Feature Engineering - 18 derived features Model Training - Random Forest + tuning Cross-Validation - 10-fold CV Evaluation - Multiple metrics

---

🔌 API Reference

Endpoint	Method	Description
/api/status	GET	Health check & API status
/api/data	GET	Dataset statistics
/api/feature_importance	GET	Model feature importance
/api/predict	POST	Purchase prediction
/api/compare_brands	POST	Brand comparison
/api/segment_analysis	GET	User segment analysis

Example Usage

import requests

response = requests.post('http://localhost:5000/api/predict', json={
    'age': 28,
    'gender': 'Male',
    'income': 65000,
    'education': "Bachelor's"
})

print(f"Purchase Probability: {response.json()['probability']}%")

---

🔍 Troubleshooting Guide

Common Issues and Solutions

Issue: "Python not found"

# Solution: Install Python from python.org
# Verify installation:
python --version
# or
python3 --version

Issue: "Module not found" errors

# Solution: Install missing dependencies
pip install flask flask-cors pandas numpy scikit-learn

# Or reinstall all requirements
pip install -r requirements.txt

Issue: "Port 5000 already in use"

# Solution 1: Kill the process using port 5000
# Windows:
netstat -ano | findstr :5000
taskkill /PID <PID_NUMBER> /F

# Linux/Mac:
lsof -ti:5000 | xargs kill -9

# Solution 2: Use a different port
# Edit Dashboard/app.py and change:
# app.run(host='0.0.0.0', port=5001, debug=True)

Issue: "Model files not found"

# The dashboard will run in demo mode
# This is normal and provides sample predictions
# Real models will be loaded if present in Models/ folder

Issue: Dashboard loads but charts don't appear

# Solution: Clear browser cache and refresh
# Or try in an incognito/private browsing window

🎯 What Users Can Do

📊 Explore Demographics

• Age Analysis: View purchase patterns across age groups
• Income Impact: See how income affects buying decisions
• Gender Insights: Compare male vs female purchasing behavior
• Education Correlation: Analyze education level impacts

🔮 Make Predictions

• Input Demographics: Age, gender, income, education
• Set Preferences: Brand, price range, feature priorities
• Get Predictions: Real-time purchase probability (0-100%)
• Understand Why: See which factors influence the prediction

🏆 Compare Brands

• Market Share: View brand popularity over time
• Loyalty Analysis: See which brands retain customers best
• Feature Comparison: Compare what each brand offers
• Demographic Preferences: See which age groups prefer which brands

💡 Gain Insights

• User Segments: Discover different customer types
• Price Sensitivity: Understand how price affects purchases
• Seasonal Trends: See when people buy phones most
• Feature Importance: Learn which features matter most

📈 Performance Notes

• Prediction Speed: < 100ms average response time
• Model Accuracy: 87.3% on test data
• Data Size: 10,000+ smartphone purchase records
• Features: 18 engineered features for prediction
• Browser Compatibility: All modern browsers supported
• Mobile Responsive: Works on phones and tablets

🚀 API Endpoints (For Developers)

When running locally, the following REST API endpoints are available:

• GET /api/status - Check API health and status
• GET /api/data - Basic dataset statistics and distributions
• GET /api/feature_importance - Model feature importance (or fallback)
• POST /api/predict - Purchase prediction for user input
• POST /api/compare_brands - Compare probabilities across brands
• GET /api/segment_analysis - Segment analysis by age, income, brand

Example API Usage:

import requests

# Get a prediction
response = requests.post('http://localhost:5000/api/predict', json={
    'age': 28,
    'gender': 'Male', 
    'income': 65000,
    'education': "Bachelor's"
})
prediction = response.json()
print(f"Purchase Probability: {prediction['probability']}%")

📚 Data Science Details

Dataset Information

• Source: Synthetic smartphone purchase behavior data
• Size: 10,841 records with 18 features
• Target: Binary classification (Will Purchase: Yes/No)
• Features: Demographics, financial, behavioral, and preference data

Machine Learning Pipeline

1. Data Preprocessing: Missing value imputation, outlier detection
2. Feature Engineering: Create derived features, categorical encoding
3. Model Training: Random Forest with hyperparameter tuning
4. Validation: 10-fold cross-validation, train-test split
5. Evaluation: Accuracy, Precision, Recall, F1-Score, ROC-AUC

Model Performance Metrics

• Overall Accuracy: 87.3%
• Precision: 85.1% (true positive rate)
• Recall: 89.7% (sensitivity)
• F1-Score: 87.3% (harmonic mean of precision/recall)
• ROC-AUC: 0.924 (excellent discrimination)

📄 Reports & Documents

• One-File Overview: Project Report/PROJECT_OVERVIEW_ALL_IN_ONE.md
• Technical Documentation: Project Report/DOCS.md
• Full ML Report (PDF): Project Report/ML Report MK.pdf

🎓 Educational Use

This project is perfect for:

• Machine Learning Students: Learn end-to-end ML project development
• Data Science Courses: Understand feature engineering and model evaluation
• Business Analytics: See how ML applies to real-world business problems
• Web Development: Learn Flask backend with interactive frontend integration

🤝 For Contributors

Development Setup

# Fork the repository on GitHub
git clone https://github.com/YOUR-USERNAME/Predictive-Modeling-for-Smartphone-Purchase-Behavior-ML.git
cd Predictive-Modeling-for-Smartphone-Purchase-Behavior-ML

# Create virtual environment (recommended)
python -m venv smartpredict_env
source smartpredict_env/bin/activate  # Linux/Mac
# or
smartpredict_env\Scripts\activate     # Windows

# Install development dependencies
pip install -r requirements.txt
pip install pytest flask-testing jupyter

# Run tests
python -m pytest

# Start development server
python run_dashboard.py

Contributing Guidelines

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Make your changes and add tests
4. Commit your changes (git commit -m 'Add some amazing feature')
5. Push to the branch (git push origin feature/amazing-feature)
6. Open a Pull Request

Areas for Contribution

• Model Enhancement: Improve prediction accuracy
• New Visualizations: Add more interactive charts
• UI/UX Improvements: Enhance user interface
• Performance Optimization: Speed up loading times
• Mobile Experience: Improve responsive design
• Documentation: Expand guides and tutorials

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

Attribution & Responsible Use

If you reuse or modify this project (commercially or non‑commercially) you must:

1. Retain the original copyright notice: Copyright (c) 2025 Mandar123454
2. Keep the MIT license text in your distribution (LICENSE file)
3. Clearly document any major modifications you make

Recommended (not legally required under MIT):

• Link back to this repository so others can find the source
• Add a short note like: “Based on SmartPredict – Smartphone Purchase Prediction Dashboard (MIT Licensed)”

No warranties are provided; see “AS IS” clause in the license. Use at your own risk.

Contribution Licensing

By submitting a Pull Request you agree your contribution is provided under the same MIT License and that you have the right to license it. If you include third‑party code, ensure its license is compatible and note it in a new THIRD_PARTY_NOTICES section or file.

Additional Legal Artifacts

The repository may include (or you can add):

• NOTICE – High‑level copyright + license pointer (for downstream bundlers)
• CONTRIBUTING.md – How to propose changes, coding standards, and review process

If these files appear later, they will complement (not replace) the MIT terms.

🙏 Acknowledgments

---

📞 Support

Need help? Here's how to get support:

1. Check the Troubleshooting Section above
2. Review Documentation: See Project Report/DOCS.md
3. GitHub Issues: Open an issue for bugs or feature requests
4. Discussions: Use GitHub Discussions for questions

---

Made with ❤️ by the SmartPredict Team
Helping businesses make data-driven decisions about smartphone market trends.