I have been engaging in analyzing sensor network data for over 10 years. One of the key techniques we have been working on has been outlier detection in data streams or time series. For many monitoring applications, identifying abnormal events is an important analytic task. Some of our studies in outlier detection from large datasets were published over 20 years ago but are still well-cited today.
Recent applications include ocean monitoring and detection of gravitational waves by laser interferometers. Both applications are not only temporal but also geospatial in nature.