Abstract
OLAP (On-Line Analytical Processing) systems provide decision makers with multidimensional analyses of large databases by generating an aggregated vision of data. Nowadays, these systems face growing non-numeric data. In this context, we propose in this paper a new generic multidimensional model called CobWeb dedicated to the OLAP of XML documents; it is based on the concept of facet. The CobWeb model aims to ease the expression of queries; also it offers an appropriate vision of the document warehouse. In this context and for manipulation purposes, we propose new visualization operators for OLAP query results by using the concept of Tag clouds as a means to help decision-makers to see the content in an efficient manner and then to focus on the knowledge in results.