OLAP and Statistical Databases: Similarities and Differences

During the 1980's there was a lot of activity in the area of Statistical Databases, focusing mostly on socio-economic type applications, such as census data, national production and consumption patterns, etc. In the 1990's the area of On-Line-Analytic Processing (OLAP) was introduced for the analysis of transaction based business data, such as retail stores transactions. Both areas deal with the representation and support of data in a multi-dimensional space. Much of the OLAP literature does not refer to the Statistical Database literature, perhaps because the connection between analyzing business data and socio-economic data is not obvious. Furthermore, there are papers published in one area or the other whose results can be applied in both application areas. In this paper, we compare the work done in these two areas. We discuss concepts used in the conceptual modeling of the data and operations over them, efficient physical organization and access methods, as well as privaccy issues. We point out the terminology used and the correspondence between terms. We identify which research aspects are emphasized in each of these areas and the reasons for that. We conclude by arguing for the support of a Statistical Object data type as one of the fundamental structures that object-oriented data models and systems should support