The first step is to reduce high dimensional data to lower dimension through Principal Component Analysis (PCA).
The second step is to access the signal- to –noise ratio in the data using project score and randomization.
The third step is to remove noise by variance filtering.The fourth step is to perform statistical test. And the last step is to use the graphs to refine the search for subgroups or clusters.
Source:http://www.genengnews.com/gen-articles/visualization-for-advanced-big-data-analysis/5947