Executive Information Systems, Inc.

Home of The New Knowledge Management, The Open Enterprise, and The EKP

Home    About EIS    Consulting    Training     Papers    Contact     Search    Professional Papers    KM Papers

 

Don't forget EIS Books and
Professional Papers!

Papers, Briefs, and Presentations On Data Warehousing, Data Mining, and OLAP

The papers, briefs, and presentations focus primarily on the above subjects and only secondarily on Knowledge Management, portal, or DKMS considerations. In instances where this distinction is not so clear, an abstract may appear both here and on the KM Papers page. The following white papers, briefs, and presentations on Data Warehousing, Data Mining, and OLAP are available.

DW Papers and DKMS Briefs and Presentations on Data Warehousing and OLAP

A Systems Approach to Dimensional Modeling in Data Marts (White Paper No. One, March 1997). This paper views dimensional modeling in data marts from the viewpoint of the Fast Analysis of Shared Multidimensional Information (FASMI) definition of OLAP. FASMI implies that dimensional data modeling supports measurement, causal, and structural modeling, as well as a specification and organization of data comprehensive enough for supporting KDD. The paper attempts to support this need by offering an approach to dimensional modeling designed to support the development of business area models of system dynamics. The approach requires highly explicit, top-down conceptualization and data inventory steps in order to sketch out the broadest possible view of the outlines of the range of measurement and cause-effect relations underlying data marts.

Data Mining and KDD: A Shifting Mosaic (White Paper No. Two, March 1997). Data Mining as a field is not yet through with the process of definition and conceptualization of the scope of the field. There are at least three distinct concepts of data mining being used by practitioners and vendors. This paper defines the three concepts, associates them with three related concepts of Knowledge Discovery in Databases (KDD), and argues that data mining is not automatic knowledge discovery, and that the dream of making it so is, at best, an ideal motivating long-term development.

Data Warehouses and Data Marts: A Dynamic View (White Paper No. Three, March 1997). This paper explores three patterns of data mart development and relationships with data warehouses: the top-down model; the bottom-up model; and the parallel development model. All three models are seen as unrealistic because they view development without explicit consideration of user feedback and its impact on development. Three related models in the presence of user feedback are then presented, their dynamics are discussed, and some predictions are made about the likely popularity of each of the three feedback models in the future.

Evaluating OLAP Alternatives (White Paper No. Four, March 1997). The rush to develop data warehouses and data marts has gained considerable momentum from the presence of server-based On-line Analytical Processing (OLAP) tools, including: Multidimensional Server-based (MDOLAP) tools; a number of Relational OLAP (or ROLAP) products; and a new tool called Sybase IQ which uses a technology we can call Vertical Technology OLAP (VTOLAP). How do we choose an OLAP product for a data warehouse or data mart? This White Paper (a) reviews the three OLAP product categories, and (b) provides a set of criteria for product evaluation in specific product contexts.

Object-Oriented Data Warehousing (White Paper No. Five, August 1997). Data warehousing has largely developed with little or no reference toObject-Oriented Software Engineering (OOSE). This is consistent with its development out of two-tier client/server relational database methodology. As data warehouses increasingly are supplemented with data marts, with data stores of diverse type and content, and with internet and intranet front-ends, the two-tier client/server paradigm has given way to a multi-tier conceptual and software framework characterized by distributed objects. Multi-tiered data warehousing needs to be reconceptualized in terms of distributed objects and therefore in terms of OOSE. This paper offers such a reconceptualization with a focus on dimensional data modeling and its relation to object modeling.

Dimensional Object Modeling (White Paper No. Seven, April 30, 1998). An object modeling approach offers advantages in supporting Dimensional Data Modeling (DDM) of data warehouses and data marts. The current approach to making the basic decisions in producing a DDM is a pragmatic one. The pragmatic approach has had considerable commercial success, but it still makes tight coupling of strategic goals and objectives to the DDM result a matter of art, rather than a product of an explicit method or procedure, results in a model composed of passive containers for data attributes, rather than components that combine both data and behavior,does not place DDM within a broader framework for integrating data and process -- that is, the pragmatic approach is too data-centric, at a time when data warehousing is concerned with integrating a complex diversity of server-based decision support system functions. This paper examines the nature of DDM and DOM, develops the argument for tight coupling of strategic goals and objectives to the DDM through an object modeling approach, and discusses the advantages of the DOM approach in more detail.

Dimensional Modeling and E-R Modeling in the Data Warehouse (White Paper No. Eight, June 22, 1998). While there is consensus in the field of data warehousing on the desirability of using DM/star schemas in developing data marts, there is an on-going controversy over the form of the data model to be used in the data warehouse. The "Inmonites" contend that the data warehouse should be developed using an E-R model. The "Kimballites" believe that the data warehouse should always be modeled using a DM/star schema. Indeed Kimball has stated that while DM/star schemas have the advantages of greater understandability and superior performance relative to E-R models, their use involves no loss of information, because any E-R model can be represented as a set of DM/star schema models. This paper discusses two issues related to the controversy. First, the claim that any E-R model can be represented as an equivalent set of DM/star schema models, and second, the question of whether an E-R structured data warehouse, absent associative entities, i.e. fact tables, is a viable concept, given recent developments in data warehousing.

Architectural Evolution in Data Warehousing (White Paper No. Eleven, July 1, 1998). This paper is concerned with DSS/data warehouse system architectural evolution in response to the growing complexity of the enterprise DSS environment and with the relationship of new architectures to a developing capability to handle the Dynamic Integration Problem. The paper briefly describes and analyzes the following architectures: Top-Down; Bottom-Up; Enterprise Data Mart (EDM); Data Stage/Data Mart (DS/DM); Distributed Data Warehouse/Data Mart (DDW/DM); Distributed Knowledge Management (DKM); Variations with introduction of the ODS. In addition it comments on the relationship between DKM architecture and data mining and provides some brief comments on software tools for implementing DKMA.

DKMS Briefs

The Corporate Information Factory or the Corporate Knowledge Factory? (DKMS Brief No. One, July 10, 1998). Bill Inmon has introduced the Corporate Information Factory. But should he have introduced the Corporate Knowledge Factory? Does it really make any difference?

Is Data Staging Relational? A Comment (DKMS Brief No. Five, November 11, 1998). A recent question raised by Ralph Kimball is whether the data staging area is relational or has more to do with sequential processing of flat files. This brief revisits and expands on Kimball's viewpoint. It examines the above issue from the viewpoint of the data staging application server, the data staging repository, and the metadata and metamodel that drive the data staging process.

Data Warehouses, Data Marts, and Data Warehousing: New Definitions and New Conceptions (DKMS Brief No. Six, November 12, 1998). Data Warehousing has lately been undergoing substantial changes in architecture and a broadening of related functional applications. With these changes have come new definitions of the Data Warehouse and evolving conceptions of data warehousing This brief explores a number of data warehouse and data mart definitions and their relation to the idea of the Distributed Knowledge Management System (DKMS). It also analyzes the meaning of "data warehousing," in light of changes in data warehousing systems and changes in definitions.

DKMA and The Data Warehouse Bus Architecture (DKMS Brief No. Seven, November 13, 1998). The Data Warehouse Bus Architecture is composed of "a master suite of conformed dimensions" and standardized definitions of facts. Business process data marts throughout an enterprise can "plug into" this bus to receive the dimension and fact tables they need. The Bus supports the various processes and associated data marts that measure key aspects of the processes. The logical union of these data marts is said to be the data warehouse. And each data mart is said to be a subset of that data warehouse. This paper describes the Data Warehouse Bus Architecture offered by Kimball, Reeves, Ross, and Thornthwaite, and then contrasts it with DKM Architecture, an object-oriented alternative.

Presentations

Architectural Evolution in Data Warehousing (September 9, 1998).

 

© 1998-2007 Executive Information Systems, Inc.