Data Profiling Challenges in Engineering Asset Management Data - Conceptual Design for Next Generation Data Profiling Software

摘要

Engineering asset management (EAM) is the process of managing the assets (from manufacturingmachines to trains, planes and road bridges etc) in an organisation. In order to manage these assets organisationsmust have good quality data about the assets. Otherwise, decisions about when to maintain an asset can be madeincorrectly, and as a consequence, can adversely impact the business financially. To improve data, the firstcommonly accepted stage is data quality assessment, and to support this stage, data profiling software is often used.Data profiling tools can be used to uncover and measure the scale of the data quality problems and they do this bydefining data quality rules. This research investigated the data profiling needs of EAM. In particular, existingprofiling tools often contain generic data quality rules that are not always applicable to EAM business users.Creating EAM data quality rules without the relevant domain knowledge is very difficult and hence the best peopleto develop these rules are the EAM business users. This research therefore proposes an enhanced data profilingsolution, which is based on the community-based central pseudo-code DQ rule repository. The proposed dataprofiling solution enables business users to develop and share EAM-related data quality rules promoting ruleadaptability and reusability.