Interdisciplinary Data Integration, Quality Assurance Study

Information systems are generally distributed in the complex information technology environment of large enterprises or organizations.Data Integration technology enables data exchange between multidisciplinary distributed systems.Upper applications,such as comprehensive data analysis,etc.,are established on integrated shared databases or data warehouses which are based on Data Integration.It is becoming more and more important that the quality of integrated data must be assured, rather than the improvement of the integration technology capabilities.Data Quality issues are prevalent in the design,development,use and integration of information systems.There are many causes for Data Quality issues,which are frequently related on imperfection of systems planning and design.The improvement of data quality requires better technology on systems design and integration.Establishing relevant business Data Standards are important means to guarantee data quality.This thesis demonstrates a comprehensive method using Data Cleansing,Data Standard,Data Auditing and Data Analysis,overall ensures the quality of Data Integration,and effectively feed back the quality issues of integrated data.Data Standards are not only adopted in the process of data cleansing and conversion,but also considered by enterprises or organizations during the unified IT planning,and adopted in distributed systems and data warehouses.This thesis refers to Dublin Core Metadata as a theoretical basis,illuminates the significant methodology for establishing a business Data Standard by an industrial case of Higher Education Management Information (HEMI).Data Quality also needs thvorable feedback mechanism.This thesis proposes Data Auditing and Data Analysis approach for discovering:and feeding back data quality issues,thereby ensures the quality of basic and integrated data in the whole process.Subsequent investigations by Data Auditing call for pertinent technical consideration on systems design,development and integration.The basic and integrated data issues could be detected during Data Analysis,solved and improved through the feedback mechanisms.