Scientific problems are rarely answered by direct comparison between A and B;
next, we want to learn about the impact of parameter 1, 2, and 3 and further,
stratify our patient groups according to shared properties and differences
observed in our data. These are constructive and necessary steps when
complex phenomena are
the objective of the studies, e.g. a high volume of multivariate biomedical
data are analyzed to give information about cause and diagnosis of a disease,
an individual‘s prognosis and potentially, the responsiveness to a special
medication.
As expected, complex phenomena are mirrored by the complexity
of the respective multi-scale, highly interconnected and condition-dependent
data. It is a real challenge to visually inspect all relevant data, decide
which level of simplification is suitable to clearly reveal patterns in the
data and points towards new insights. Especially, multidimensional data
patterns are difficult to recognize when encoded into two dimensions.
Therefore, the real artistry lies in the extraction of relevant information
by applicable analytical methods and the explorative visualization of the results.