Data modeling for improving performance of data mart

Abstract

Data warehouse is a technology of potentially enormous worth to an enterprise. It is defined as a subject oriented, integrated, time-variant, and nonvolatile collection of data in support of the management decision-making process. Data mart is basically another form of a data warehouse, but it is a decentralized subset of data found in the data warehouse. Although there are many pros and cons about the usefulness of the data mart, it must be a powerful, rapidly constructed, tool for delivering business information. Instead of using traditional data model in order to build a data mart on the enterprise Intranet, we present another way of modeling the data mart, which allows only one record for each customer to improve processing performance although many transactions for a customer might be happening over time. To do so, we have to do some operations, e.g., number crunching on the fields in each record. We then develop a real world application doing data mining tasks using the data mart.