🍊 Citrus Quality Classification - Machine Learning Project

A comprehensive machine learning project that classifies orange quality into three categories (Good, Medium, Poor) based on physical, chemical, and environmental characteristics. Features a trained logistic regression model with 98.6% accuracy and an interactive web application.

📋 Project Overview

This school project demonstrates an end-to-end machine learning pipeline from data analysis to deployment. The system helps farmers and distributors automatically grade orange quality for better market placement and pricing decisions.

🎯 Features

• 🤖 Machine Learning Model: Logistic Regression with 98.6% cross-validation accuracy
• 🛠️ Complete Pipeline: Integrated data preprocessing, training, and evaluation
• 🌐 Web Application: Interactive Streamlit app for real-time predictions
• 📊 Data Visualization: Comprehensive EDA and performance metrics
• 🔧 Production Ready: Model persistence and easy deployment

📁 Dataset

500 samples with the following features:

Feature	Description	Type
diameter	Orange diameter (cm)	Numerical
berat	Weight (grams)	Numerical
tebal_kulit	Skin thickness (cm)	Numerical
kadar_gula	Sugar content (%)	Numerical
asal_daerah	Origin region	Categorical
warna	Skin color	Ordinal
musim_panen	Harvest season	Categorical
kualitas	Quality label (Target)	Categorical

Quality Classes:

• 🟢 Bagus (Good) - Export quality
• 🟡 Sedang (Medium) - Local market quality
• 🔴 Jelek (Poor) - Industrial processing quality

🏗️ Model Architecture

Pipeline([
    ('preprocessing', ColumnTransformer([
        ('scaler', StandardScaler(), numeric_features),
        ('ohe', OneHotEncoder(), categorical_features)
    ])),
    ('model', LogisticRegression())
])

📈 Performance

Metric	Score
Cross-validation Accuracy	98.6%
Test Accuracy	100%
Precision	99%
Recall	99%

🚀 Quick Start

Prerequisites

Python 3.8+
pip install -r requirements.txt

Installation

1. Clone the repository

git clone https://github.com/yourusername/citrus-quality-classification.git
cd citrus-quality-classification

2. Install dependencies

pip install -r requirements.txt

3. Run the Streamlit app

streamlit run app_jeruk.py

Usage Examples

Train the model:

from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split

# Load and preprocess data
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
model = LogisticRegression()
model.fit(X_train, y_train)

Make predictions:

# Single prediction
new_orange = [[7.5, 180.0, 0.5, 12.0, 'Jawa Barat', 'oranye', 'kemarau']]
prediction = model.predict(new_orange)
probability = model.predict_proba(new_orange)

🖥️ Web Application

The Streamlit app provides an intuitive interface for quality prediction:

Features:

• Interactive sliders for numerical features
• Pill selectors for categorical options
• Real-time quality predictions
• Probability visualization
• Business recommendations

📂 Project Structure

citrus-quality-classification/
│
├── app_streamlit.py                 # Streamlit web application
├── model_klasifikasi_jeruk.joblib # Trained model file
├── jeruk_balance_500.csv        # Dataset
├── requirements.txt             # Dependencies
├── EDA_analysis.ipynb          # Exploratory Data Analysis
└── README.md                   # Project documentation

🛠️ Technical Stack

• Programming: Python 3.8+
• Machine Learning: Scikit-learn, Pandas, NumPy
• Visualization: Matplotlib, Seaborn
• Web Framework: Streamlit
• Model Persistence: Joblib

📊 Results Analysis

Feature Importance

The model identifies key factors affecting orange quality:

1. Sugar content (kadar_gula)
2. Weight (berat)
3. Skin thickness (tebal_kulit)
4. Diameter (diameter)

Business Impact

• Farmers: Better pricing decisions based on quality
• Distributors: Optimal market channel selection
• Exporters: Automated quality control for international standards

🔮 Future Enhancements

• Compare multiple algorithms (Random Forest, SVM, Neural Networks)
• Add feature importance analysis with SHAP values
• Develop REST API for integration
• Mobile app development
• Real-time image recognition for quality assessment

👥 Contributors

• Your Name - GitHub Profile
• School: SMKN 1 Purbalingga
• Course: Machine Learning

📄 License

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

🙏 Acknowledgments

• Dataset provided for educational purposes
• Instructors and peers for valuable feedback
• Open-source community for amazing libraries

---

⭐ If you find this project useful, please give it a star!

---

📞 Contact

For questions or collaborations, feel free to reach out:

• Email: ibraramadanialzaki2@gmail.com
• LinkedIn: Your Profile
• Portfolio: Your Website

---

Made with 🍊 and ❤️ for Machine Learning