TiP: Analyzing Periodic Time Series Patterns

Time series of sensor databases and scientific time series often consist of periodic patterns. Examples can be found in environmental analysis, where repeated measurements of climatic attributes like temperature, humidity or barometric pressure are taken. Depending on season-specific meteorological influences, curves of consecutive days can be strongly related to each other, whereas days of different seasons show different characteristics. Analyzing such phenomena could be very valuable for many application domains. Convenient similarity models that support similarity queries and mining based on periodic patterns are realized in the framework TiP, which provides methods for the comparison of similarity query results based on different threshold-based feature extraction methods. In this demonstration, we present the visual and analytical methods of TiP of detecting and evaluating periodic patterns in time series using the example of environmental data.