Abstract. It has been well recognized that online analytical processing (OLAP) can provide important insights into huge archives of data. While the conventional OLAP model is capable of analyzing relational business data, it often cannot fit many kinds of complex data in emerging applications, such as bio-medical data, time series and semi-structured data. In this paper, we propose GOLAP, a general OLAP model. We show that GOLAP is consistent with the conventional OLAP model on multidimensional databases. Moreover, we show that the model can be applied to complex data as well. As an example, we illustrate a research prototype system, GeneXplorer, which enables OLAP over gene expression data.