Data Lifecycle Process Model and Quality Improving Framework for TDQM Practices

Abstract

Total Data Quality Management (TDQM) methodology claims that we should manage data as we manage products, but principles, guidelines, and techniques of Total Quality Management for physical products have limitations in TDQM practices because of the nature of data. Some concepts and models are proposed for resolving problems in TDQM practices. Firstly, history evolvement of product is discussed, and data will be the sign of information age just as maching products, electric products, electron products, and software products do. So, many advanced theories and methods being used for improving other products' quality coud be introduced into TDQM. Secondly, data lifecycle process model facing data analysis is presented. Data's lifecycle process is divided into three phases, design, manufacturing and application. Based on the model, quality problems are resolved in the phase which they arise initially and avoided complication in the next operation. Problems could be found and resolved as early as possible. Then, two technological means for improving data quality, data cleaning and data integration, are analysed and compared. Last, a data quality improving framwork conforming to TDQM for data acquisition flow with hiberarchy is proposed, and data quality may be controled by background process without any impact on regular application in foregrounding.