🚨 Fraud Detection System

---

📌 Project Overview

This project is an end-to-end Machine Learning application that detects fraudulent financial transactions using a Random Forest Classifier.

Built with Python and Streamlit, the application predicts whether a transaction is fraudulent, provides a fraud probability score, and presents the result through an intuitive dashboard designed for real-time decision-making.

---

🌐 Live Demo

🔗 https://fraud-detection-system-xhtapccyai4u7weahuzfzt.streamlit.app/

Note: This application is hosted on the free Streamlit Community Cloud. If the application is asleep, simply click "Yes, get this app back up!" It usually starts within a minute.

---

🎯 Business Problem

Financial fraud results in significant losses for banks, payment providers, and businesses worldwide.

The objective of this project is to assist in identifying suspicious transactions early, allowing organizations to reduce financial losses and improve fraud prevention through machine learning.

---

🤖 Machine Learning Model

Item	Value
Algorithm	Random Forest Classifier
Problem Type	Binary Classification
Target Variable	Fraud (0 = Safe, 1 = Fraud)
Class Imbalance	Class Weighting

---

📊 Dataset

This project uses the Kaggle Credit Card Fraud Detection Dataset.

https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud

Dataset characteristics:

• Real-world financial transactions
• Highly imbalanced dataset
• Anonymous features (V1–V28)
• Suitable for fraud detection research

---

✨ Features

• ✅ Real-time fraud prediction
• 📈 Fraud probability score
• 📊 Risk visualization
• 🌐 Interactive Streamlit dashboard
• ⚡ Instant predictions
• 🤖 Machine Learning powered

---

🛠️ Technology Stack

• Python
• Streamlit
• Pandas
• NumPy
• Scikit-learn
• Matplotlib
• Joblib

---

📂 Project Structure

fraud-detection-system
│
├── app
│   └── app.py
│
├── data
│   └── creditcard.csv
│
├── models
│   ├── model.pkl
│   └── features.pkl
│
├── screenshots
│   ├── Home.png
│   ├── SafeTransaction.png
│   ├── SafeTransactionchart.png
│   ├── FraudDetected.png
│   └── FraudDetectedchart.png
│
├── train_model.py
├── requirements.txt
└── README.md

---

📸 Application Preview

🏠 Home Screen

---

✅ Safe Transaction Prediction

---

🚨 Fraud Detection

---

💼 Business Impact

This solution demonstrates how machine learning can support financial institutions by:

• Detecting suspicious transactions earlier
• Reducing potential financial losses
• Supporting faster fraud investigations
• Improving operational decision-making

---

🚀 Future Improvements

• Add SHAP explainability
• Compare Random Forest with XGBoost
• Deploy a FastAPI backend
• Add email notifications
• Improve dashboard analytics

---

👨‍💻 Author

Craig Chiambiro

🎓 BSc Information Technology (Data Science)

💼 Software Developer | Data Analyst | Machine Learning Enthusiast

🔗 GitHub: https://github.com/CraigChiambiro

💼 LinkedIn: https://www.linkedin.com/in/craig-chiambiro-6b3394257

---

⭐ If you found this project useful, consider giving it a star!