🚨 Factory Guard AI

AI-Powered Industrial Safety Risk Prediction System

A production-oriented Machine Learning system designed to proactively identify unsafe industrial conditions and reduce workplace accidents through intelligent risk prediction.

---

📌 Problem Statement

Industrial workplaces face significant safety challenges due to:

• Lack of real-time monitoring
• Human error and unsafe practices
• Delayed identification of hazardous conditions

Traditional systems are reactive, responding only after incidents occur.

🎯 Objective

To build a proactive AI-driven safety system that predicts potential risks and enables early intervention.

---

🧠 System Overview

Factory Guard AI is an end-to-end Machine Learning application that:

• Ingests industrial condition parameters
• Applies a trained LightGBM model
• Predicts risk levels (Safe / Unsafe)
• Displays results through an interactive Streamlit dashboard

🔁 High-Level Workflow

1. Data Collection
2. Data Preprocessing
3. Feature Engineering
4. Model Training (LightGBM)
5. Model Evaluation
6. Deployment via Streamlit UI

---

🏗️ System Architecture

[Input Data] → [Preprocessing] → [LightGBM Model] → [Prediction] → [Streamlit UI]

---

⚙️ Tech Stack

Layer	Technology
Language	Python
ML Framework	LightGBM
Data Processing	Pandas, NumPy
Visualization	Matplotlib, Seaborn
Web Interface	Streamlit
Model Persistence	Pickle
Version Control	Git & GitHub

---

📊 Model Details

🔍 Algorithm: LightGBM

A gradient boosting framework optimized for speed and performance.

✅ Why LightGBM?

• Handles large-scale structured data efficiently
• Faster training compared to XGBoost
• High accuracy with lower memory usage

📈 Performance Metrics

Metric	Score
Accuracy	~98%
F1 Score	~0.87
ROC-AUC	~0.98

🧪 Training Pipeline

• Data Cleaning
• Feature Scaling
• Train-Test Split
• Model Training
• Hyperparameter Tuning
• Evaluation

---

📸 Screenshots

🔹 Application Interface

🔹 User Input & Predictio

🔹 Output Result

🔹 Performance Visualization

---

🚀 Getting Started

🔧 Prerequisites

• Python 3.8+
• pip

---

📥 Installation

git clone https://github.com/your-username/FACTORY-GUARD-AI.git
cd FACTORY-GUARD-AI
pip install -r requirements.txt

---

▶️ Run the Application

streamlit run app.py

---

🌐 Access

Open in browser:

http://localhost:8501

---

📁 Project Structure

FACTORY-GUARD-AI/
│
├── app.py                # Streamlit application
├── train_model.py       # Model training script
├── requirements.txt     # Dependencies
├── README.md            # Documentation
├── screenshots/         # Project images
└── model.pkl            # Trained ML model

---

📈 Results & Impact

• Achieves high prediction accuracy (~98%)
• Enables early detection of unsafe conditions
• Reduces dependency on manual monitoring
• Can significantly improve industrial safety compliance

---

🔮 Future Scope

• 🎥 Integration with Computer Vision (CCTV monitoring)
• 📡 Real-time data streaming (IoT sensors)
• 📲 Alert system (SMS / Email / Dashboard notifications)
• ☁️ Cloud deployment (AWS / Azure / GCP)
• 🤖 Deep Learning models for complex pattern detection

---

🧪 Use Cases

• Manufacturing Plants
• Construction Sites
• Chemical Industries
• Smart Factories (Industry 4.0)

---

👨‍💻 Author

Mangasamudram Lokeswari

Data Science | Machine Learning | Generative AI Enthusiast

GitHub: https://github.com/Lokeswari3173

---

⭐ Support

If you find this project useful:

• ⭐ Star this repository
• 🍴 Fork it
• 📢 Share with others

---

📜 License

This project is open-source and available under the MIT License.

out put screenshorts