Modeling and processing of time interval data for data-driven decision support

Over the past decades, several disciplines like artificial intelligence, music, medicine, ergonomics or cognitive science dealt with problems concerning analyses of data associated with time intervals. Topics like pattern recognition, comparison, quality, or visualization are in focus of current research. Using these techniques in the context of data-driven decision support is quite rare even though the importance of data to support better decision making can be enormous. Reasons lie above all in limited insufficient tooling support, expensive data processing, and inapplicable requirements. In this paper, we discuss the use of time interval data and name difficulties arising when processing such data for data-driven decision support. We discuss and present solutions for overcoming the identified problems and enabling the usage of time interval data for data-driven decision support. We introduce a time interval data analysis model that provides fast access to the raw time interval data but especially to aggregated time series, mostly needed when making meaningful decisions.