Looking at CAREY: Controlling Climatological Regions in State of Emergency

We present CAREY, an alert system notification for regions in state of emergency. CAREY implements the mediator-wrapper architecture on top of the Geospatial Web to visualize risky regions in terms of their weather conditions. We illustrate the formalization of the problem of detecting risky regions as a two-fold problem that relies on annotations of sensor data with the ontology of Observations and Measurements (O&M) to enhance state-of-the-art data mining and ranking techniques. First, sensor observations are clustered according to their Geospatial information then, proximate regions are clustered into micro-climate areas in terms of the similarity of their weather conditions. Finally, Top-k Skyline techniques are used to identify the top-k areas that best meet criteria of risk among the areas that are incomparable with respect to this condition. We demonstrate the capabilities of CAREY.