Live Telemetry Demo

Live Classification, Anomaly Detection, and Dashboard Reporting

Send controlled test events to the Cloudflare Worker and review Gradient Boosting classifications, Isolation Forest anomaly results, final Worker risk decisions, and offline model evaluation evidence in the Streamlit Dashboard.

Live Telemetry Test

The buttons below generate controlled project-demo events. Each event sends minimized telemetry to the Cloudflare Worker. The Worker validates the event, forwards the approved 14-feature schema to the Azure Flask prediction API, applies the final risk logic, and stores the result in Cloudflare D1.

Beacon Result

No beacon event sent yet.

Privacy note: test events do not include raw IP addresses, names, emails, passwords, cookies, authentication tokens, private content, exact location, full URLs, or invasive browser fingerprinting.

Dashboard and Model Evaluation

The Streamlit Dashboard is the reporting layer for recently stored telemetry, supervised classifications, unsupervised anomaly results, Worker risk decisions, and offline model evaluation evidence. The current prototype provides advisory actions and does not automatically block traffic.

Live Prototype Summary

Recent events loaded, high-risk events, average risk score, detected anomalies, and unique page categories from the latest Worker event window.

Live Detection Results

Gradient Boosting traffic classifications, Isolation Forest anomaly results, final Worker risk levels, and detailed telemetry records stored in D1.

Offline Model Evaluation

Separate holdout evaluation results for Gradient Boosting and Isolation Forest, including accuracy, balanced accuracy, precision, recall, F1-score, false-positive rates, anomaly metrics, and confusion matrices.