------------------------------------------------------ URA 2055 ---------------- Author(s) : THIOULOUSE J. Towards Better Graphics for Multivariate Analysis: the Interactive Factor Map Computational Statistics Serial Number : 11, 1996, 11-21 Key-Words : Multivariate analysis, exploratory data analysis, interactive graphics, factor map, principal component analysis. Abstract The interpretation of the factors computed by multivariate analysis methods is often a difficult task. The main difficulties come from the cluttering of the factor map when the data set is large, and from the need to visualize the data when interpreting the factors. An interactive method based on the representation of the data on the factor maps is presented, with reference to a computer program available on the Internet network. Several techniques offered by this program are demonstrated with the help of a real-size example data set: searching, zooming, and plotting of rows and columns of the data table on the factor map. E-Mail : Jean.Thioulouse@biomserv.univ-lyon1.fr Laboratoire de Biometrie, Genetique et Biologie des Populations (URA 2055) - Univ. C. Bernard LYON I - 69622 VILLEURBANNE CEDEX