Spatio-temporal data handling and multivariate statistics
Data useful for precision decisions and operations are commonly spatio-temporal of nature. Our data are generally multivariate, typical for data orginating from multi- and hyperspectral instruments. We utilize both existing and self-developed solutions to handle such data. For standardized and consistent storage, we mainly use open geo-data standards and database systems. To extract useful information, we apply state-of-the-art multivariate statistics. The use of powered partial least square regressions (PPLS) in combination with a range of pre-processing algorithms has proven to be a particular strong tool for such data mining.