How yosefshabaan/cats-online-anomaly-detection works
This repository combines a notebook-based autoencoder anomaly-detection experiment, documented Kafka streaming simulation, and a Streamlit app for exploring model results and running detection on uploaded sensor CSVs. The app routes sidebar selections to individual pages; execution validates data, loads a selected model and threshold, and visualizes anomalies. Evaluation compares saved experiment reports. The notebook and Kafka workflow are described in the README and illustrated as code in the Colab page; the notebook itself was not sampled, so its internal stages are not expanded.
Main entry points: app.py App Router · execution_page.py Execution Page · AnomalyDetection.ipynb Training and Streaming · evaluation_page.py Evaluation Page
Components 17
Notebook Workflow
- Training and Streaming
AnomalyDetection.ipynb
Streamlit App
- App Router
Streamlit_result_visualization/app.py - Home Page
Streamlit_result_visualization/menu_pages/home_page.py - Project Info
Streamlit_result_visualization/menu_pages/model_page.py - Colab Guide
Streamlit_result_visualization/menu_pages/colab_page.py - Evaluation Page
Streamlit_result_visualization/menu_pages/evaluation_page.py - Execution Page
Streamlit_result_visualization/menu_pages/execution_page.py - Contact Page
Streamlit_result_visualization/menu_pages/contact_page.py
Detection Runtime
- CSV Validation
Streamlit_result_visualization/menu_pages/execution_page.py - Model Selection
Streamlit_result_visualization/menu_pages/execution_page.py - Anomaly Scoring
Streamlit_result_visualization/menu_pages/execution_page.py - Anomaly Visualization
Streamlit_result_visualization/menu_pages/execution_page.py
Models and Results
- Model Weights
Streamlit_result_visualization/models - Evaluation Reports
Streamlit_result_visualization/models_info
Other components
- User
- Sensor Dataset
- Kafka
Connections 16
- User → App Router: selects page
Streamlit_result_visualization/app.py - App Router → Home Page: routes to
Streamlit_result_visualization/app.py - App Router → Project Info: routes to
Streamlit_result_visualization/app.py - App Router → Colab Guide: routes to
Streamlit_result_visualization/app.py - App Router → Evaluation Page: routes to
Streamlit_result_visualization/app.py - App Router → Execution Page: routes to
Streamlit_result_visualization/app.py - App Router → Contact Page: routes to
Streamlit_result_visualization/app.py - User → Execution Page: uploads CSV
Streamlit_result_visualization/menu_pages/execution_page.py - Execution Page → CSV Validation: validates with
Streamlit_result_visualization/menu_pages/execution_page.py - Execution Page → Model Selection: builds and loads
Streamlit_result_visualization/menu_pages/execution_page.py - Model Selection → Model Weights: loads weights
Streamlit_result_visualization/menu_pages/execution_page.py - Execution Page → Anomaly Scoring: runs inference
Streamlit_result_visualization/menu_pages/execution_page.py - Execution Page → Anomaly Visualization: plots results
Streamlit_result_visualization/menu_pages/execution_page.py - Evaluation Page → Evaluation Reports: reads reports
Streamlit_result_visualization/menu_pages/evaluation_page.py - Sensor Dataset → Training and Streaming: provides time series
README.md - Training and Streaming → Kafka: simulates streaming
README.md