Practicalities in applying a formal methodology to data analysis

Abstract

This paper draws on the practical experience of CACI consultants in Europe in database design and data analysis over the past three years. It is not intended to explain in detail or to justify the particular methodology, but rather to concentrate on the practical problems that arise when undertaking data analysis and the solutions which we have found to be successful. Further, the emphasis of this paper is on the technical problems, whereas it is the political and organizational aspects which pose the greatest difficulties for the data analyst.

The methodology employed by CACI is similar to those described in other papers presented at this conference, in that it is based on representing an environment in terms of entities and their attributes and relationships. It originated from work done by Ed Tozer and myself during 1973 and was subsequently extensively enhanced by David Gradwell and myself and latterly by contributions from a number of CACI database consultants.

The data analysis methodology when used in preparation for database design involves two distinct phases — entity analysis and functional analysis. The inherent nature of the environment and its representation in the form of data, is analysed distinctly (although not independently) from an analysis of the business activities and the uses to which data is put. Details of both the data and its processing are essential to the design of an effective database where system performance is a consideration. However, this paper deals largely with entity analysis and does not give the same attention to the practicalities of functional analysis.