🚨 Real-Time Anomaly Detection Dashboard (NAB)

A portfolio-ready project demonstrating real-time anomaly detection on the Numenta Anomaly Benchmark (NAB) dataset.

This project simulates industrial sensor anomaly detection with a live Streamlit dashboard for monitoring.

A portfolio-ready project demonstrating real-time anomaly detection on the Numenta Anomaly Benchmark (NAB) dataset.

This project simulates industrial sensor anomaly detection with a live Streamlit dashboard for monitoring.

---

🎯 Why This Project

In Industrial AI, early detection of anomalies in machine or sensor data can:

• Prevent costly equipment failures ⚙️
• Improve worker and system safety 🦺
• Enable predictive maintenance 🔧
• Optimize industrial processes 📈

This project demonstrates how real-time monitoring systems can be built using machine learning, time-series data, and dashboards.

---

📊 Demo

Dashboard in Action

(If the GIF doesn’t load, check it directly in the repo: assets/demo.gif)

---

🚀 Features

• 🔄 Simulated streaming of NAB datasets
• 🧠 Anomaly detection with Isolation Forest (swap with PyOD/deep learning)
• 📈 Interactive dashboard with live charts & controls
• 🟥 Visual anomaly highlighting (red dots on chart + red rows in table)
• 📊 Evaluation metrics (Precision, Recall, F1 updated live)

---

⚡ Quickstart

# 1) Clone your repo, then set up env
python -m venv venv
source venv/bin/activate  # Windows: venv\Scripts\activate
pip install -r requirements.txt

# 2) Get NAB dataset (official repo)
git clone https://github.com/numenta/NAB.git

# 3) Run the dashboard
streamlit run app.py