Sensor Anomaly Visualization Engine





Image:SAVE user interface visualizing sensor network topology (top-left, bottom-left), high-dimensional sensor measurements and statuses (top-right), dimension correlations (middle of bottom) and the dimension temporal trends (bottom-right). Various visual settings and analytic functionalities can be accessed from the menu.




1. Develope several novel visualizations based on generic anomaly detection methods, for effective representation of the dynamic high-dimensional sensor data

2. Provide a new way for operators and administrators to identify sensor network data patterns, diagnose potential failures, and locate root causes of these failures in a user-friendly and interactive manner

3. Prototype and evaluate the above techniques



·        Lei Shi, Qi Liao and Yuan He,SAVE: Sensor Anomaly Visualization Engine, IEEE Conference on Visual Analytics Science and Technology (part of IEEE VIS), October 2011. [paper] [code]