Analyzing dynamical data - from basic statistics to modern time series analysis
given by Dr. Reik Donner
Reik Donner studied physics and mathematics at the University of Potsdam, with a PhD in physics for a thesis on multivariate analysis methods for palaeoclimate time series awarded in 2007. After postdoctoral positions at the TU Dresden, the Max Planck Institutes for Physics of Complex Systems (Dresden) and Biogeochemistry (Jena) and the Potsdam Institute for Climate Impact Research (PIK), he became leader of a BMBF-funded Young Investigators Group focusing on the development of complex systems methods for understanding causes and consequences of past, present and future climate changes at PIK in 2014. He had been awarded a Guest Professorship at the Osaka Prefecture University in Sakai, Japan, in 2009, and an Outstanding Young Scientist Award of the European Geosciences Union (EGU) Division "Nonlinear Processes in Geophysics" in 2011.
When it comes to analyzing empirical observations or model output data, many researchers commonly resort to basic statistical tools, potentially missing a whole world of detailed information on underlying processes that these simple methods cannot resolve by their construction. This course provides an introduction into the world beyond these classical statistical methods. Starting from corresponding basic analysis tools like correlation functions, common problems are introduced that arise when dealing with real-world time series across scientific disciplines (like handling nonstationarity, trends, periodic components and stochastic persistence), together with mathematical approaches addressing the corresponding conceptual challenges. Besides discussion the underlying concepts and their limitations, a particular focus will be on providing particular examples on how to use these data analysis techniques in common statistical software environments like R or Matlab, and how to interpret the thus obtained results.
