Missing Data in Interactive High-Dimensional Data Visualization
Deborah F. Swayne, Andreas Buja · 1998
this paper we documented potential uses of linked brushing (as in figure 2) to explore missing value patterns and associations between missing values and the variables of interest. Other interactive data visualization methods can be equally useful and apply immediately in this approach. We hope to have provided a useful adjunct to traditional methodology for missing value problems. We expect it to be quite useful for data exploration as well as for imputation diagnostics. The XGobi software is freely available from StatLib at this URL: http://lib.stat.cmu.edu/general/XGobi/ For more references on XGobi, visit the following (identical) web pages: http://www.public.iastate.edu/~dicook/xgobi.html http://www.research.att.com/~andreas/xgobi.html References