I built this interactive demo to visualize 'data drift' in industrial sensors. It simulates a motor's vibration data in the browser. When you click 'Loose Bolt', it injects noise that increases the Coefficient of Variation without necessarily spiking the mean immediately. The backend is Python (Flask + Numba) calculating a stability score on a sliding window of the last 30 points. Thought it was a fun way to visualize a dry statistical concept. Let me know what you think!