Exploratory data analysis (EDA) plays an increasingly important role i
n statistical analysis. EDA is difficult, however, even with the help
of modern statistical software. We have developed an assistant for dat
a exploration, based on Al planning techniques, that addresses some of
the strategic shortcomings of conventional software. This paper descr
ibes the design and behavior of the system and discusses an experiment
al evaluation that demonstrates the effectiveness of our approach. (C)
1998 Elsevier Science B.V. All rights reserved.